[Pkg-javascript-commits] [dojo] annotated tag 1.5.3 created (now f452de0)

David Prévot taffit at moszumanska.debian.org
Thu Aug 21 17:39:01 UTC 2014


This is an automated email from the git hooks/post-receive script.

taffit pushed a change to annotated tag 1.5.3
in repository dojo.

        at  f452de0   (tag)
   tagging  1708abccb0cbc9269b85ea78f8fe309152e38e3d (commit)
 tagged by  Colin Snover
        on  Wed Aug 21 22:34:29 2013 -0700

- Log -----------------------------------------------------------------
Release 1.5.3

Adam Peller (380):
      dojo.i18n and tests. Refs #2624
      load dojo.i18n from bootstrap. Refs #2624
      unit test for dojo.i18n. Refs #2624
      Oops... locale setting was already in hostenv_browser.  Undo.  Refs #2624
      Port ant script to build cldr localization data. Refs #2624
      Fix annoying String.replace problem on IE and pass unit tests.  Refs #2624.
      Refs #2704
      Refs #2704
      Not thrilled with asserts - will revisit later.  Also, still need to port io/xhrGet code when ready.  Refs #2704
      Refs #2704
      Refs #2704
      Refs #2704
      First stab at porting Animation and lfx (fx).  Refs #2710
      Refs #2710
      Fix regression in fade animation.  Refs #2710
      The first test in JsonItemStore (fetch_all) fails for me.  Jared, can you take a look?  Refs #2717
      Still trying to get unit tests working.  Refs #2710
      Update to use dojo.connect.  Refs #2710
      Port dojo.number and currency along with string and regexp dependencies.  Refs #2719
      Fix slideIn/slideOut. Refs #2710
      Port unit tests for number and currency. Refs #2719
      Check in common set of cldr transforms, as built by xslt files.  Fixes #2719, #2325, #2326, #2327
      port dojo.lfx.slideTo, some style changes and reductions. Refs #2710
      fix dojo. reference.  Refs #2727
      Generate cldr currency data without objects.  Fixes #2742
      Fixes #2749
      Add dojo.fx.wipeIn, add some tests. animate test not yet working. Refs #2710
      Fix NaN tests. Refs #2719
      Refs #2710
      Missed a dojo.html reference. Refs #2710
      more fixing up for 0.9. Refs #2710
      fix wipeOut.  Refs #2710
      wipeIn/Out was renamed to slideIn/Out. Refs #2710
      Fixes #2769
      Initial checkin of dojo.date.  Needs more cleanup for 0.9M2.  Refs #2758
      Changed selector API for dojo.date. Added dojo.cldr UT.  Refs #2758
      Fix unit test references so they all run.  Refs #2758
      format returns null for invalid input. Refs #2860
      Handle parse of empty string. Fixes #2930
      Part 2 of date refactor, to be continued.  Basde on Owen's suggestions. Refs #2758
      Get the dojo.date unit tests running again. Refs #2758
      apply dojo.date refactor to dijit/dojox. Refs #2758
      oops. didn't mean to check in loader.  Refs #2758
      Replace enum-style args with string literals.  Refs #2758
      Fixes #2350
      More dojo.date refactoring based on Owen's suggestions. Refs #2745
      Date refactor -- move helper functions out of dojo.date.  Refs #2758
      Provide a bit more flexibility to allow null references to be run through the format function, as intended.  Refs #2350.
      Inline set*OfWeek methods, change paths to get*OfWeek methods to conform to new package conventions.  Refs #2758
      Thanks, Doug.  Missed some regexp special chars.  Fixes #3102
      Fix millisecond problem on IE by removing timestamp-style arguments, plus some style fixes.Fixes #3112
      Replace for loops with dojo.forEach to reduce code.  Fix non-strict comparison for month parsing.  Fixes #2729
      Group space between currency symbol and number with the symbol in the regexp.  Fixes #3036
      Fixes IE regression. Fixes #2729
      added some tabs to make Alex happy :-P  dojo.cldr is mostly generated data and therefore really doesn't have any standalone functional tests; dojo.{date,number,currency} ought to include tests which use this data.  Refs #2758
      Implement options.fractional, set to [true,false] by default for currencies.  Fixes #2862, #2860
      Forgot to port README for CLDR. Refs #2719.
      Refs #3220
      Replace Rfc3339/Iso8601 methods with to/fromISOString methods to resemble ECMAScript 4 spec.  Fixes #3007
      Additional comments. Refs #3007
      make iso date unit tests pass with new API. Refs #3007
      tests.date.stamp UT was failing on IE.  Refs #3007
      Use faster regexp methods.  Refs #3007
      Assume local timezone if none is specified in fromISOString. Fixes #3386
      Improve docs.  Fixes #3338
      Remove instance-based chain/combine methods, fix event connections on combine.  Update affected Dijit code.  Fixes #3325.
      don't need to log these squelches. They're legit. Refs #3453
      Break out dojo.Color into its own module.  Fixes #3343
      missing 'var'.  Fixes #3575
      Fixes #3605
      Replace removeChild with dojo._destroyElement. Refs #1927
      Provide translations for dojo.colors.  Use mixins to add CSS3 to existing colors in _base.  Fixes #3188
      Add translation comments, fix a few spellings. Refs #3188.
      more spelling fixes.  hope I got them all this time. Refs #3188
      More efficient color conversion, plus inline docs. Fixes #3667
      Fixes #3699
      Bogus checkin to test post-commit script. Refs #3699
      Bogus checkin to test post-commit script. Refs #3699
      Bogus checkin to test post-commit script. Refs #3699
      Bogus checkin to test post-commit script. Refs #3699
      Bogus checkin to test post-commit script. Refs #3699
      Patch to workaround BiDi problems calculating offsets on IE.  Fixes #3115
      Forgot RTL test file. Refs #3115
      Don't try to unload bundles in teardown to restore state.  delete isn't sufficient, and we don't do this for any other modules.  Fixes #3605
      Fixed rgb2hex so that it returns 3 octets, not 4.  But I still don't think animateProperty is working right for colors. Fixes #3667
      Fixes #3651
      Fixes #3699
      IE doesn't support iframe.contentDocument.  Refs #3720
      Ugh.  We were off by a scrollbar. Fixed logic in getIeDocumentElementOffset (thanks, Justin)  Refs #3720
      Put a more descriptive throw in _Templated for missing keys.  Don't want to just return an empty string, or the developer may never notice the error.  Also allow user to use HTML literals in templates using "!" prefix. Fixes #3700
      replace another trim func. Refs #3731
      Go back to iterative approach to calculate day/week of year, due to DST bug. Fixes #2587, #2024
      dojo.date.difference also fixes #2587, #2024
      Fix issue #1 in currency test. Refs #3672
      Fix issue no. 1 in currency test for numbers and dates. Refs #3672
      Use dojo.fromJson for default case, rather than eval() for clarity and re-use.  Refs #2727
      Account for the fact that IE doesn't treat nbsp as whitespace in regexps.  Refs #3672
      am/pm can come before or after the hour.  Fixes #3672.
      Do whitespace replacement within [] regexp.  Fixes #3837
      Minor style, spelling changes.  Refs #2556
      Move source uri hint to end of file (for firebug 1.1) so that it won't throw off the line count in the eval buffer by 1.  Refs #3092
      Remove bogus char(s).  Refs #3938
      Rename slideIn/slideOut to wipeIn/wipeOut.  Fixes #3917
      Rename slideIn/slideOut to wipeIn/wipeOut.  Refs #3917
      Remove accidental global reference.  Refs #4066
      Fix parse/regexp with places ranging from 0 to n: "0,n".  Fixes #2864, Refs #3906
      Initial checkin of OpenAjax hub, r100 (not yet integrated into bootstrap) under Apache License 2.0.  Refs #4145
      pick up a couple more bytes from array.  Use xor bitwise logic on booleans.  Refs #3961
      Comply with Dojo style.  Refs #4145
      Use new docparser comment feature for 'faux declaration'.  Refs #3961
      Fix typo I made in 'faux declaration' of getComputedStyle.  Refs #3961
      Add comments for isObject plus style mods.  Fixes #4424, #3961
      Handle "class" and "style" attributes according to IE quirks. Refs #3058
      Style and space fix for dojo.isArrayLike. Refs #3961
      era was broken. Fixes #4548
      Translations for dojo.color. Refs #2389
      Fix up German color translations. Refs #2389
      Reductions for _base, including:
      Fix failure log statement.  Fixes #4722
      Tolerate no space before am/pm in parse.  Fixes #4718
      some code reductions. Refs #3961.  Throw if json filtered comment fails.
      ship a copy of the unicode license with modified ldml. Refs #2719
      Just do the registerLibrary call, for now.  Fixes #4145
      Handle false in parseResults.  Fixes #4744
      include license text. Refs #2872
      it's not a problem with date, and the validation tests pass. strangely this bug seems limited to DateTextBox in Forms?  Refs #4760
      sorry, didn't understand the original bug report.  How's this? (probably close to where we started)  Fixes #4744
      Perhaps this is slightly better. Refs #4744
      In a build, an exact match is required for locale.  Fixes #4814
      Fix regexp for json comment filtering to accept multilines.  Fixes #4829
      Test for unmatched substitute. Refs #4763
      When formatting time only, do not include a timezone.  Fixes #4879
      Rename NodeList extension to reflect fx additions.  Refs #4480
      I give up. Put back the indexOf searches for comments in place of regexp attempt.  Refs #3961, #4829, #4888  Reverts part of change in [10834]
      Fix bugs in _containsValue - wasn't searching the whole list.  Use dojo.some.  use toString() for regexp match on primitives.  Refs #4931.  Still needs UT.
      oops. remove commented out code. Refs #4931
      Bryan's patch to fix DST problems. Refs #5047
      Fixes really old regression in add weekday routine.  Thanks, Bryan for catching this.  Refs #5053
      updated translations. Refs #5184
      Add final Greek strings.  Thanks again, Dionysios.  Fixes #5205
      Fix typo. Fixes #5291
      Move singleton docs inline, where possible.  Fill in some missing summaries.  Refs #982
      Fill in missing summary.  Refs #982
      xml response handler - can't return out of a dojo.forEach, and prefixes was undefined.  Fixes #5320  Also, make objectToQuery shorter and faster.
      Fix doc typos. Refs #4820
      put onload trigger in try/catch so a failure doesn't kill the bootstrap. Refs #5197
      repair encoding regression in [11732] and add test cases.  Fixes #5415
      Fix unresolved reference 'url'. Refs #5415
      Fix xhr unit test for json pp change.  Also, reduce json code some more.  Refs #5421
      Updated test to explicitly check for failures with Boolean coercion - Refs #5200.  Invalid month/day names on parse now return null. Thanks, Doug.  Fixes #4864
      so it seems right to revert this to [10980] unless there's a reason not to.  Fixes #5464 !strict
      Some code reductions. Fixes #5375.
      put global funcs in local scope. fix b0rken array logging.  Fixes #5489  !strict
      Correct DnD/dropdown bidi placement.  Fixes #4369, #4228.  Also some size reductions. !strict
      Fixes bidi in quirks mode (thanks, makin) Fixes #3721 and backs out change from [11944] Refs #5509. Plus some size reductions.  !strict
      GB weeks start on Monday. Fixes #5533
      some reductions (=== undefined) and style mods.  Refs #5542
      oops.  default for date should be 1, since dates are 1-index based.  Refs #5595
      isDisabledDate should compare date portion only.  DST bug?  Fixes #5595 !strict
      build results of new cldr 1.5.1.  Fixes #4503
      Treat 2 year dates as 4 year dates, regardless of the pattern for strict=false.  Refs #4850
      Update copyrights to 2008.  Thanks to doughays for reminding me :)  Refs #2607
      fix lowercase map to pass i18n/date unit tests. Refs #4503
      build the Date object at the end, not as we go along.  The setters on Date seem to have a bunch of issues.  Calling the Date constructor once is more reliable and probably more performant.  Also, use 1970 as the default year.  Fixes #5716, #5680
      reduce the bounds checking code. Refs #5716
      another small reduction. Refs #5716
      remove deprecated argument juggling in declare.js.  style fixes. !strict  Refs #3961
      Fix regression from [12198] - must convert hours to a number for comparison.  Refs #5716
      throw for bad json. Refs #5725
      Unnecessary to check year overflow in parse.  Refs #4850
      Always format as 4-digit year in dijit.form.DateTextBox, but continue to allow 2-digit year input where appropriate.  Refs #5465 !strict
      Use regexp to find cookie contents. (thanks, Pat) Fixes #2881.  Correct docs on deleting cookies.  Fixes #5782 !strict
      Missed reference in reduction. Refs #2881. !strict
      Reduce toPixelValue with a regexp, remove redundant return statements. Refs #3961 !strict
      Remove wildcards in toStyleValue regexp.  Still just as accurate as it was before. Refs #3961 !strict
      One more time... don't assume pos=0 Refs #3961 !strict
      Don't need to seed cache with width/height. Refs #3961 !strict
      remove useObject, since it's really a one-liner for the developer.  Fixes #5810. reduce some code.  use expires to remove test cookie. Refs #4969  !strict
      Refs #5885
      Fix typo, fixes #5897. Thanks tvachon
      Years < 100 are not supported.  Pad years < 1000 when formatting.  Fixes #5751
      Fix up comments to match signatures, show multiple types on arguments and document their behavior, use common naming of parameters.  Refs #4820
      Fix IE crash (thanks Mike Wilcox), change deprecated/experimental to console.warn.  Fixes #5490  Refs #5922 !strict
      Apply Mike Wilcox's firebug patch.  7 up.  Fixes issues with Safari, popup blocker, resizing. Adds cookie support to remember position and size.  Fixes #5922 !strict
      Typo. Fixes #5946
      Better, but still not great, definition for dojo.data. Refs #4820
      Add Wolfram's test case for string callbacks.  Fixes #5905.  Had to modify exception test since error string seems to vary across browsers.
      Incorporating fix, but can't take test case w/o cla. Refs #5793
      Avoid infinite loop in indexOf. Fixes #5961
      Ok, it really was JavaScript 1.6, despite the misleading URLs.  Refs #4820
      add more descriptive error message for onload failure. Refs #5197
      Use markdown, kwargs in docs. Refs #5962
      Fix strict mode checking of year < 100CE in dojo.date.parse. Refs #4850.  Also remove unsupported whitespace tests and invalid 'es' locale tests. Fixes #5885
      Correct kwArgs doc declaration, plus some other doc fixups.  !strict Refs #5962 and r12790
      Doc function param. Refs #5962
      Improve docs and markdown. Refs #5962
      Improve docs and markdown. Refs #5962
      Check the day of the month of the PST-defined date in UTC time, since the local day will vary with timezone.  Fixes #6079
      Append //@ to buffer for debugging only for Gecko-based browsers. Fixes #5901
      Add CurrencyTextBox constraints docs.  Refs #6075 !strict
      Fix currency constraints docs.  Refs #6075 !strict
      pad milliseconds field. Fixes #6118
      fix cookie props doc. Refs #5962 !strict
      Add hypertext link to docs.  Refs #5962 !strict
      Improve djConfig docs.  Refs #5962
      Clean up docs. Refs #5962 !strict
      Clean up docs. Refs #5962 !strict
      Clean up docs. Refs #5962 !strict
      Clean up docs. Refs #5962 !strict
      Clean up docs. Refs #5962 !strict
      Clean up docs. Refs #5962 !strict
      Honor round option in format(). Fixes #6279
      Group 1 and group 2 translations for Dojo 1.1.0.  Refs #6314
      Return undefined for JSON requests with HTTP Status 204, rather than throw, since it's successful by definition.  Fixes #6601  Fix typo in dojo.query comments and style of assignments in conditionals. !strict
      Thanks for the great patch, Seth!  Fix regexps for ddMMMyyyy date parse.  Fixes #6242
      Move Norwegian resources from 'no' to 'nb', to be compliant with latest ISO
      Refs #6793.  Doc typo, remove extra wording.
      Good catch.  Fix case to match 'standAlone'.  Refs #6801
      Handle patterns which start with a literal at char 0.  Fixes #6915  Merci, Matthieu et Damien!
      Remove broken 'e' pattern char and unspecified 'L' and 'c'.  Refs #6660
      Update dojo.cldr to use new generation scripts in [14022].  Fixes regressions seen with CLDR 1.5.  Fixes #6813
      Fix regression caused by [14023] so month names show up in dijit._Calendar again.  Fixes #7041
      Complete the comment. Refs #7041
      Avoid rounding issue in dojo.date.difference by comparing Date objects with the same hour.  Fixes #6960
      Use UTCMinutes for timezones +:30  Fixes #7014
      Redo logic to fake aliases. [14130] failed in a build. Refs #7041
      Make it clear that dojo.fromJson throws exceptions. Fixes #7082
      Fixes #6744. Really just a problem with the comparison.  419.3 should be Safari2, anything greater Safari3 (for now).  Thanks, jgarfield.
      Add mapping for 'he' locale to assume 'il' for country.  Fixes #7126
      Update to CLDR 1.6.  Regenerate dojo.cldr.  Fixes #7113
      use undefined for falsey results from dojo.isXxxx methods like dojo.isIE, instead of 0.  Fixes #7159, #7153
      simplify IE8 check.  Refs #7159 !strict
      Update tests to work with CLDR 1.6.  Provide loose parsing of 'a.m.'. Refs #7113.
      Add docs for dojo.xhr args.  Only problem left is that IOArgs load, error, and handle property docs aren't showing up -- is function signature confusing jsdoc?  Refs #6793
      Try to improve wording (please review) and add some TODOCs. Refs #6793
      oops, put back response arg docs.  But can't response be an Error object or structured data?  Don't we need to be more descriptive?  Is there any point to using the function(...){} syntax, since it's not parsed and only makes this less clear? Refs #6793
      Allow ranges "n,m" for number formatting, for symmetry and to fix Spinner widget.  Fixes #4477
      number parsing allows unlimited places when pattern is not provided explicitly.  Fixes #6536
      Honor fractional:false for dojo.currency.format. Fixes #7091
      get a couple of bytes back.  Refs #7480, [14917] !strict
      Avoid instanceof check on Element for prototype.js compatibility.  Fixes #7467 !strict
      Run browser-based tests conditionally.  Refs #7523, #7178
      Add ca, sk, sl, th translations, plus all translations for PasswordValidator.  From IBM (CCLA) Fixes #7481
      update dojo.cldr with improved xslt scripts to follow aliases.  Fixes #7188
      Use arbitrary subset of currencies by default. Refs #7188
      Correct iphone example. Refs #982
      Update dojo.cldr with proper aliasing to preserve abbreviations.  Fixes #7607
      Change dojo._loadModule documentation to reference public API dojo.require.  Refs #6505
      add a comma, remove a comma.  Refs #7461
      Fix docs for num arg on rep. Refs #7461
      Fill in example for dojo.string.pad. Refs #7461
      Update dojo.cldr.  Fixes #7665, Refs #7666, #7849.
      Ignore am/pm with alt attributes. Refs #7849
      Implement dojo.date.locale.displayPattern.  Fixes #7666
      code reductions. Refs #3961
      make nameAnonFunc a local function inside the closure. Refs #3961
      Use some instead of forEach searching for ActiveXObjs (thanks, Alex)  Improve corresponding test. Refs #5388
      Remove query dependency on dojox.dom method.  Fixes #5052
      Translation updates from Lotus Domino for sl/sk.  Refs #7481
      use String.indexOf   Refs #7666 !strict
      better. thanks, cougar.   Refs #7666 !strict
      Put in comment, since all non-extension FF devs will now hit the try/catch.  Fix up some minor style issues.  Refs #3833 !strict
      fix some doc typos.  Refs #7461 !strict
      Improve algorithm to try to round out binary floating point errors, and document it.  Fixes #7930
      Normalize whitespace/nbsp handling for currency.  Refs #8036
      oops. remove console.log. Refs #8036
      oops. remove console.log. Refs #8036
      Add isFunction unit test.  Refs #7845
      generalize _getQueryFunc for Chrome.  Refs #7845 !strict
      Style things.  Refs #7845 !strict
      Event normalization should include Chrome.  Style fixups also.  Refs #7845 !strict
      Remove unnecessary chars from escapeString expression.  Refs #8109
      Style cleanups.  Refs #7390 !strict
      remove platform is* initializations; assume undefined. Refs #9999 because I can't get to trac.
      Trim out a few bytes. Refs #8137
      Accounts for ugly IE bugs in toFixed().  Refs #8078
      Fix typo.  Fixes #8181 !strict
      Save a couple of bytes.  Refs #7942 !strict
      Assure that round goes away from zero on ties.  Improve docs.  Refs #7930, #8036
      Tolerate leading zeros in whole part without separators. Fixes #6933
      revert isFunction() check for isWebKit.  Unit test added in [15738].  Refs #7845
      Use 0.9 in our test instead of 0.5, so we'll be less susceptible to other browser bugs.  Refs #8280
      Add dojo.cldr locales to svn to match translated list.  Fixes #8342
      Escape "+" in regexp literals and add tests.  Fixes #8346
      Use native String.trim() where available.  Fixes #8182
      Update docs to indicate native String.trim() is used.  Refs #8182
      Remove unneeded call to str.trim().  Refs #8182
      Check for setTimeout and clearTimeout definitions explicitly, since the reference does not seem to throw in Rhino 1.7.  Fixes #8159 !strict  Thanks, Mark Wubben!
      Additional patch to correct a Rhino bug with bidi override chars in the build.  like dojo.fromJson, dojo.requireLocalization also uses eval on the file contents and must be patched.  Fixes #3808 !strict
      should have hit save first.  Include cast to JS String.  Refs #3808 !strict
      Update minor rev in lieu of script. Refs #8428.
      remove hostenv comment inserted in [16348]  Refs #8428
      Undo inlining of rev.  Distributed SCM like git and bzr provide alternate access to our svn and will not substitute keywords.  Refs #8428
      save a few bytes. Refs #8367
      save a few more bytes. Refs #8367
      reduce some more and add unit tests. Refs #8367
      dojo._base.window is browser-dependent.  Avoid tests in non-browser environments.  Refs #8367
      Change console.debug call to console.error, refs #2924, fixes #8583.  (Do we really need a console message here at all if the error can be detected programatically?  Style change refs #7250
      doc typos !strict refs #7461
      Support 14 sig figs.  Fixes #8699, Refs #7930 !strict
      Make sure Infinity/-Infinity are formatted to null.  Refs #8756
      move floating point fudge from dojo.number.round to dojox.math.round (experimental)  Refs #7930, Refs #8756, Refs #8699
      Remove test(s) that refs #7930, while I pursue what appears to be a Firefox bug.
      specify a locale on all culturally-sensitive tests. Refs #9076
      specify a locale on all culturally-sensitive tests. Refs #9076
      Cause unknown interval for dojo.date.add to throw.  Code reductions. Fixes #8973
      Fixes IE event connect issue on trunk, refs #9093.  !strict Thanks, jfcunat
      Checking in translation updates into trunk.  Updates all locales except ca, th, sk, sl.  Fixes #7691
      !strict Fix typo. Refs #4824
      Style and doc updates. Refs #5962
      Use 1am for date-timese where midnight is a dst leap.  Fixes #9366, Refs #8521 !strict
      Merge [17730] to correct comment.  Refs #9366
      One more tweak.  DST could occur on the first of the month.  Refs #9366
      Inject "//@ sourceURL" debugging hints for all browsers, except IE, where it might conflict with conditional compilation.  Fixes #9374.
      corrections to Simplified Chinese. Refs #7691
      check in dojo.cldr build result for CLDR 1.7.1.  Fixes #9445
      Remove patternChars method, since it's been deprecated in the CLDR.  Fixes #9489  Use substitution pattern for date-times.  Use dateFormatItem for year-only patterns.
      remove tests referencing displayPattern method.  Refs #9489
      Update to cldr 1.7.1  Refs #9445
      Try to fix jsdoc for dojo.version.  Refs #8606
      translation updates. Refs #7481
      Expand aliases in non-gregorian calendars.  Refs #9043
      Remove islamic-civil from bundle list, since it seems to be causing errors.  Fixes Buddhist month names.  Refs #9043
      Test update from Yoav Rubin (IBM) to reflect draggable areas.  Refs #9593
      Land Rob Retchless' patch.  Refs #7925 Will follow up with some sort of DOH tests, if possible.
      Update from Rob Retchless, IBM.  Refs #7925
      Remove hash tests from module.  Halts test suite. Refs #7925
      put hash tests back.  Refs #7925
      Docs should properly reflect that scientific notation is not yet supported.  Refs #9126
      build datetime strings from cldr pattern.  Refs #9445
      Add branch for year < 100, apparently used by .net.  Fixes #9281
      Call decodeURIComponent again, tests pass.  Fixes #7925
      Check for exceptions if connection fails in IE.  Patch from retchless.  Fixes #10176 !strict
      new dojo.hash tests. thanks, miksago!  With tweak from retchless. Fixes #10003, refs #10336 !strict
      use dojoBlankHtmlUrl  Fixes #10379  Thanks, rcoup !strict
      Fix currency property inline docs.  Refs #8606
      Fix up currency/number constraint mixin docs.  Refs #8606
      fix typo in docs. Refs #8606 !strict
      Leave encoding/decoding up to user.  Patch from retchless.  Fixes #10595, Refs #7925 !strict
      correct handling of double single-quotes. Fixes #10554.  Thanks, toonetown
      Uncomment test cases that no longer fail. Refs #10554
      Set locale to 'en' on test. Good catch, youngho! Refs #10554
      Move 'fractional' property definition from currency to number.  Refs #8606
      more description of dojo.currency.  Refs #8606
      more description of dojo.currency.  Refs #8606
      syntax error in docs. Refs #8606
      one more time. Refs #8606
      Schema change for CLDR 1.8.  Refs #10692
      Manually update these files to reflect changes in CLDR 1.8.  Refs #10692
      Ran script to update dojo.cldr.nls based on CLDR 1.8.  Refs #10692, #10479
      Modify property name for am/pm strings (Refs #10692) list hebrew leap month as separate property and modify Dijit to handle leap month (Fixes #9044) !strict
      Oops.  Missed a checkin.  Refs #10692, Fixes #10909.  Thanks, youngho.
      Fix tests.  Refs #10692
      add comment to reflect that literals are not supported in number patterns. Refs #10843
      Remove empty dirs. Refs #10479
      Process field/relative tags.  Refs #7986, #10120
      typos. Refs #8606
      make defaultTime work on IE.  Good catch, jbondc. Fixes #10969
      test defaultTime. Refs #10969
      Allow decimal places when parsing numbers type=percent.  Fixes #10701, Refs #10582
      Fixes #10669 - formatting a number with constraints.places indicating a range
      Remove stray global \!strict  Fixes #11032.  Thanks, brettz9
      Redo r21792 to avoid exception when parts[1] is null and options.places is unspecified.  Refs #10701, Refs #10582
      Update to CLDR 1.8.1-prelim.  Fixes Chinese am/pm.  Fixes #10692
      Strip out location-dependent timezone from date format test.  Refs #10692  Add number percent test.  Refs #10701, #10582
      update copyrights to 2010  \!strict  Refs #2608
      Fix a long-standing regression on IE8, brought about by r21776.  Thanks, wildbill, for noticing!  Refs #3007
      Ignore Timezone prefixes like +09:00  Fixes #11198
      security warning. Refs #10694
      re parses left-to-right, fix day expression to properly capture 30, 31.  Thanks, echtykov.  Fixes #11332
      declare bogus Date("x") to clear up Safari5 NaN bug. \!strict Refs #11279.
      Another attempt to ignore timezone prefixes.  Refs #11198 Fixes #11409
      Adding Romanian translations  Refs #10470
      Using r22487 to create a branch named 1.5.
      Remove unicode bidi control chars, 1.5 branch.  Fixes #10847 \!strict
      Kazakh translations Refs #11648
      patch from retchless \!strict fixes #11257
      Updated translations. Fixes #11648
      Applying Bryan's patch to fix am/pm overrides.  Fixes #12155
      translation corrections.  Refs #12322

Alex Russell (476):
      git-svn-id: http://svn.dojotoolkit.org/src/dojo/trunk@7417 560b804f-0ae3-0310-86f3-f6aa0a117693
      refs #2550
      adding the HTML runner harness. Updating the base test harness.
      implements async test harness. Refs #2550
      adding getTestCallback() method as discussed with Brian Skinner today in IRC (already doc'd in the porting guide). Almost have the test URL story figured out. Refs #2550
      updating with Brian Skinner's excellent suggestions. Refs #2550
      changing the names to assert* from is*. Refs #2550
      un-eff the browser harness in IE. Refs #2550
      ensuring that logging output shows up in the log tab. Refs #2550
      adding directories for basic Base APIs and platform specializations. Refs #2500
      pulling in the throwaway bootstrap file. Refs #2500
      as per private email from James. Refs #2500
      porting in the bootstrap for 0.9 modification. Refs #2500
      refs #2500
      refs #2500
      refs #2500
      refs #2500
      refs #2500
      updating the throw-away bootstrapper to reference the new locations of the loader files. Refs #2500
      port the command-line tests to run using the most primordial Dojo package loader. Refs #2500. Refs #2550
      keep the test system from conflicting w/ Dojo when run in a browser environment. Refs #2550
      implements changes that allow both the Rhino and Spidermonkey hostenv's to run the test harness smoke test successfully. Refs #2500
      fixes onload code, removes TONS of cruft, and slims the built (browser) bootstrap loader to 11K from 21K. Refs #2500
      provide console.* stubs for environments that don't have Firebug. Refs #2500
      ensure that all tests get loaded in from tests/_base.js. Refs #2550
      ensure that the console is cleared for each test run. Refs #2550
      ensure that the play/pause button is in the right state. Refs #2550
      dojo.post_load_ --> dojo._postLoad. Refs #2500
      dojo.debug --> console.debug. Refs #2500
      remove dojo.findModule. It was effing useless. Refs #2500
      prevent leakage of an inadvertant global name. Refs #2500
      go straight for the document.all case in IE. Refs #2500
      ensure that the base loader get added to the mix. Refs #2500
      pull in stuff that the Base needs. Refs #2500
      beginning port of Deferreds. Refs #2500
      refs #2609
      initial start at pulling in some of the utility funcs. Refs #2608. Refs #2609
      refs #2608. refs #2609
      ugg. It's late. Refs #2609
      nowhere near working yet. Refs #2610
      try harder to get function-only test names, add a minimal test file for the bootstrap, and ensure that test harness self-tests aren't run all the time. Refs #2500
      pull in the query system to Base
      pull in the query() system to Core
      better conformance w/ the refactor document. Refs #2609
      refs #2610
      adding basic array tests. Still need to add test for lastIndexOf. Refs #2610
      removing unused functions. Refs #2610
      note to self. Refs #2500
      add tests for Deferred and fix Deffered to not use dojo.raise(). Refs #2610. Refs #2609
      removing last arg to getObject() after discussion w/ jburke. Refs #2500
      from discussion w/ james burke. Refs #2500
      ensure that the array methods work correctly on IE/Opera/Safari/etc. Earlier revs were using the removed dj_global and is* functions to tell type. Refs #2610
      add NodeList and query to the Base. Refs #2501
      start logging error objects directly. Helps us get actual failure info out of exceptions. Huzzah. Refs #2550
      brings NodeList and query() up to date. I'm surprised at how small the diff is. Still need to bring in/write tests for them. Refs #2615
      fix test-name-finding on Opera. Refs #2550
      get debugging output working correctly on the webkit nightlies. Refs #2550
      output how many tests need to be run. Refs #2550
      tests and updates to the basic lang utilities. Refs #2609
      fixing a bug and adding more array tests. Refs #2610
      bogus checkin to confirm that #2552 is solved. Refs #2552
      adding docs to base loader. Refs #2500
      fixing style, updating comments, and un-breaking toJson. Refs #2609
      updating docs, expanding description, fixing style. Refs #2609
      special-case NodeList in isArray() and use the slightly faster comparators for isFunction. Refs #2609
      fix loading on IE. Refs #2609
      completing test set for array utilities. Updating utils to use fast-path if available. Fixes #2610.
      refs #2597
      enable the djConfig attribute on <script> tags that pull in Dojo. Simplifies setting config options. Refs #2500
      fixes requireIf's for rhino testing. Refs #2501
      make sure that we don't bomb out in non-browser envs. Refs #2609
      dojo.raise() is removed in 0.9. Refs #2624
      refs #2615
      include query() tests. Refs #2615
      adds support for:
      keep the system from throwing errors when a test page that includes dojo but doesn't have a parent harness loads it. We still get success/failure info at the console. Refs #2550
      more fixes to prevent errors with test pages not hosted in harness. Refs #2550
      adding dojo.Uri and dojo.moduleUri. Also removing the "walk up" for .* packages. Refs #2500
      tests and correctness fixes for dojo.query(). This is something of a backslide on performance, but correctness trumps. Still need to find another week or so to spend on query(). Refs #2615.
      adding sounds so that success and failure sound like...well...success or failure. Refs #2550
      beginning of tests for the NodeList class. Refs #2615
      we probably don't want selfTest to always be on when not loading Dojo. Refs #2550
      ensure that we actually play the "woohoo!" sound if it all goes well. Rate limits sounds. Refs #2550
      more correctness fixes for non-xpath code paths. Refs #2615.
      creating t, f, and is aliases for assertTrue, assertFalse, and assertEqual (respectively). I'm getting really freaking sick of typing the long names. Refs #2550
      a basic summarizer. Refs #2550
      seems the cleanup was throwing errors. Fixes getObject() tests on IE. Refs #2500
      ensure that we dont' fail when loaded stand-alone. Refs #2550
      prevent us from bombing out should we be loaded stand-alone. Refs #2550
      adding a first cut at style, DOM, and html utils. Refs #2686
      ensure that coords() works at least minimally. Refs #2686
      updating style code w/ the APIs Scott and I had agreed on. Updating the NodeList to use the new (corrected) APIs. Still need to implement style reading and setting. Refs #2686. Refs #2615
      adding quirks-mode tests for HTML namespace. Refs #2686
      command line tests were b0rken. This fixes. Refs #2500
      enable a bit of debugging output for rhino and sm. Refs #2500
      porting in dojo.declare() and tests for it. Huge thanks to sjmiles and Steve from TurboAjax for their contribution of this code. Fixes #2692
      IE byId hack only for IE < 7, get/set style properties, and a fix for boxMode calculations typo. Refs #2686. Still needs docs and tests.
      ensure that object comparison tests can end successfully. Refs #2550
      minor updates to address logging issues on Opera and IE. Refs #2550
      great catch by Jared Jurkiewicz on async test completion w/ sync results. Refs #2550
      dumb error. Thanks again to Jared for catching it. Refs #2550
      un-busting NodeList's filter method. Refs #2615
      minor change to let root's be IDs. Refs #2615
      only branch for setOpacity on IE. Refs #2686
      formatting fixes. Correctly detect firefox and IE versions now. Refs #2500
      un-fscking NodeList.filter(). Push was pushing *all* the args. Refs #2615
      removing async options for _getText. It was completely unused. Refs #2500
      fixes byId issue on IE. Refs #2500
      fixing strange issues on IE. Refs #2686
      IE is a steaming pile. Refs #2686
      correctly scope queries. Not sure how I'd missed this. Guess we need more tests for query(). Refs #2615
      some reformatting of the fx code to remove use of the dojo._base.fx object. Also enabling it in the default require list for Base. Refs #2710
      ensure that we get non-pixel-valued styles without casting and ensure that we do it in a compact and relatively performant way. Refs #2686
      ensure that the tests anticipate IE b0rkennes. Fixes more IE6 bone-headedness. Ensures that setting opacity values in IE returns the right value. Refs #2686
      Opera reports computed color styles in #xxxxxx format. Test accordingly. Refs #2686
      unfscking log output for Webkit and Opera. Refs #2550
      make the log output slightly smaller. Refs #2550
      use a more accureate success tests. Also make semi-private to allow the xhr* functions to use it directly. Refs #2500
      minor style nits. Refs #2609
      API location refacotoring, style updates, and reformatting. Refs #2641
      ensure that paused groups resume correctly. Refs #2550
      first cut of Ajax core + tests. Refs #2711
      implement rawXhrPost, add tests for xhrPost(). Refs #2711
      code re-ordering to survive the build. M1 tests now pass post-build. Refs #2615. Obsoletes the attached patch.
      potential fix for sync requests. Refs #2716. Needs verification
      make sure the xhr tests get added to the default set. Refs #2711
      refs #2418
      give this test more breathing room. Refs #2710
      implement NodeList.style() and NodeList.styles(). Refs #2686
      adds a delay between test execution. Refs #2721. May fixt it, but it'll need more testing and Jared's verification.
      making the console() tests pass on Safari. Refs #2500
      making the default CSS file readable. Refs #2597
      make body text readable. Refs #2597
      make safari/ff work right on OS X. Refs #2597
      use our default CSS for the test runner. Refs #2597
      use a font that's more available on XP. Refs #2597
      starting to make comments readable. Still need more work on this front. Refs #2749
      more formatting cleanup. Refs #2749
      more NodeList tests. Refs #2615
      make our test pages look like dojo. Refs #2597
      ensuring that we exercise all the filter() argument sets. Refs #2615
      super-trivial diff. Refs #2704
      now that we've got lots of test groups, roll them up but make them expandable.  Refs #2550
      changing the returns of NodeList's style() and styles() functions. Returning a NodeList from them is idiotic. Refs #2615
      IE has "issues" with calls to apply() on the COM XHR object. Work around is to directly call into open(). Fixes #2767
      committing Scott's new keyboard handling work + tests. Refs #2641. Still needs more unit tests.
      I'm a freaking idiot. My apologies to scott and adam for screwing things up for both of them. Refs #2641
      fixing runner accuracy in reporting group failure. Refs #2791
      moving sounds to the new test dir. Refs #2795
      more movement. Refs #2795
      moving the browser runner. Refs #2795
      final DOH files. Refs #2795
      ensure that "doh" has a default mapping (to the anon/committer view layout) in the package system. Refs #2795
      tests.* -> doh.* migration. Moving DOH out of Core. Fixes #2795
      more tweaking to the defaults. Force IE to cache background CSS images to prevent request flooding. No need to make things unreadably small, 13px is just fine. Refs #2597
      removing the Moz x86_64 workaround. FF 1.5 is nearly EOL'd anyway. Refs #2711
      adding Ben Lowery's patch for dojo.cookie(). Nice work, Ben. Fixes #2796
      ensure that textareas are included in form serialization. Refs #2711
      setting native EOL on _base files. Refs #2850
      ensure native EOL. Refs #2850
      native EOL. Refs #2850
      native EOL. Refs #2850
      native EOL. Refs #2850
      native EOLs. Refs #2850
      fix DnD on IE. Refs #2728
      fixing formatting for comments. C'mon people, if it doesn't look readable to you, it's not. Refs #2719
      roll in the YUI CSS rules now that we've vetted the IP. Should speed up page loading considerably. Fixes #2872.
      make sure our test pages look like Dojo test pages. Refs #2597
      fixes #2757
      break stuff up into individual test fixtures to assist in debugging. Refs #2757
      fixes #2913
      A modified, dojo-ized, style-guide-compliant version of Firebug Lite for use in
      Thanks to Scott Miles for fighting for these methods.
      style guide conformance for changeset 8402. Refs #2641
      use indexOf instead of regexp's. Bed-lumping _docScroll. Refs #2686
      dumb mistake. Refs #2686
      another dumb mistake on my part. Refs #2686
      patch from Ben Lowery for numeric additions to cookies. Refs #2796
      and updated version of html.js from Scott Miles. Passes unit tests and improves performance. Thanks to him for pushing so hard on me to get this right. Refs #2686
      cleanup for bed-lumped names. Refs #2686
      merging the first part James' patch to add ScriptSrc IO to Core. Refs #2924
      merging James' patches (source and tests) for script src IO. Fixes #2911
      fixes the tests. Refs #2924
      test stub for RPC. Refs #2915
      more cleanup from Scott's back-porting. Refs #2686
      fixes weirdness w/ nodes that have display="none" on Safari. Refs #2686. Helps to fix a slew of widget-related breakage on shipping Safari.
      fixing XHR Post regression added in last rev. Refs #2924
      fixes issues w/ context nodes + "> ..." rules. Fixes #2615. New issues on query should be filed under separate cover.
      should fix content serialization bug. Brings things closer to style-guide compliance. Refs #2915. Tests are still failing, though.
      make sure that the NodeList tests don't bring up the firebug list console when fired up stand-alone. We don't want to screw the results. Refs #2615
      animation isn't useful if we can't manipulate colors. Adds dojo.Color class (completely reworked) and adds support for color animations back into the the Base animation system. Wish it were smaller, but it is what it is. Refs #2710
      starting port of dojo.behavior to 0.9. Refs #3001
      start of work to port to 0.9. Refs #3001
      make it possible to declare this store from markup w/ the parser. Fixes #3021
      ben's inital port. Refs #2914
      merging a (great) patch from peller. Fixes #3033
      adding tests for dojo.behavior.
      remove spurious debugging. Refs #2710
      move the regexes out of the fast path so that there's less chance they'll be recompiled. Drops heavy dojo._Url usage speed by ~20%. Refs #3092
      adding test todo file. Refs #3121
      make sure things get zindex more correctly on Webkit, Opera, and IE. Refs #3140
      give the fx unit tests 15 seconds to finish. They were timing out intermittently before. Fixes #3141
      un-effing <em> and <strong> tags so that dojo.css can actually be used in RichText areas. Refs #3010
      logging some oversights in testing. Refs #3121
      thanks to david schontzler for coming out of retirement to spot some errors in html.js. Refs #2686
      adding "json-comment-optional" handleAs type. Refs #2520
      fixes #3194. Makes forEach() tollerent of b0rken array params just like the built-in variant. This seems to fix rendering on themeTest.html in which the checkbox widget was passing a null array to forEach.
      adding Google Gears detection support. Refs #3252
      ensure that repeated showings of a dialog continue to show the background. Improve background show/hide to actually fade things in and out. We want dialogs to be visually attractive. More work on transitions may be in order (e.g., Lightbox effects). Refs #3247
      post, put, and delete support from Jared. Fixes #3276
      ensure that return order is right on IE/Safari. Fixes #3293
      add a connect() method to NodeList so we don't have to do so much forEach()-ing to set up a behavior. Refs #2615
      fixing more of the Java-in-JavaScript cruft that pervades the offline code. Ensuring that the _loadedUrls property of the dojo object actually acts as an array. Refs #3252
      beginning of move of parser into core. Refs #3365
      updates for stubs to improve compat. Refs #3365
      some terminology changes to ensure that things "feel" as general as they are. Refs #3365
      make query() more resiliant in whitespace parsing for compund queries. Refs #2423
      first cut of a <script type="dojo/connect"> system. Refs #3387
      source -> event as per discussion w/ Owen. Refs #3387
      thanks to Cameron Braid for spotting this typo. Refs #3387
      bed-lumping the event connection machinery in the parser. Refs #3193
      make sure attribute searches handle all-lower-case variants on IE. Not true case insensitivity, but it's as close as we're gonna get on that browser. Fixes #3438
      make sure we're logging errors. Refs #3453
      provide animation for node lists. Refs #2615
      was missing a passthrough arg. Refs #2615
      ensure that connect() works correctly in non-browser hostenvs. Fixes #3530
      tests for dojo.parser. Refs #3535
      make sure we log our squelches. Refs #3453
      adding first-argument preamble plucking. Makes impedence matching easier when you don't want to muge the prototoype's preamble. Refs #3535
      updating the parser to:
      test for markupFactory(). Refs #3535
      making it look a bit better. Refs #3052
      adding dojo._destroyElement method. Fixes #2931
      implement djConfig.parseOnLoad param. Move tests to use parseOnLoad to signify that they want/need full page parsing onload. Lots of hygene fixes for the dijit test pages. Fixes #3510
      adding inline docs for html functions. Fixes #3162
      adding isDescendant() and setSelectable() methods. Refs #3121. Still need to update dijit to use the new methods.
      turn parsing on in the parser tests ;-). Refs #3682
      ensuring the that parser tests pass and that the parser is working for some builtins attrs where it previously failed. Fixes #3682
      fixes #78. The name of the ".*" files is now settable via djConfig.packageFileName, which defaults to "__package__". A global, recursive rename is required on the part of users affected by #78, but that's not our issue. mod_rewrite or similar provide a way out.
      style nits. comments and docs need fixing badly. Fixes #1837
      ensure that XML documents are passed back unserialized on IE, even when it's being lame. Fixes #3294
      accepting Eugene's clone() patch. Fixes #3553
      adding a toggler class and the start of unit tests for the dojo.fx namespace. Fixes #3641. Refs #3121
      slight variable factoring for size. Refs #3961
      adding better in-code docs for xhr.js. Fixes #3975
      Prevent in-flight leakage on IE. fixes #2357
      adding docs for many of the NodeList methods, adding tests for some of them, improving the code structure in general to make it clear that extending NodeList is straightforward, and expanding the unit tests to check for return types from NodeList methods. Still more work to be done there, though. Refs #3244
      adding developer docs for dojo.query(). Fixes #3908
      fixes to NodeList.map and eliding away the node list single arg ctor method to match IE's (b0rken) behavior. themeTester.html should work on IE now. Fixes #3987
      ensure that all unit tests are being run. Fixes #4060
      migrate the rather tedious <script type="dojo/method" mode="connect" event="....">  syntax to simply <script type="dojo/connect" event="...">. Fixes #3615
      use parseFloat instead of parseInt where appropriate. Fixes #4061
      Refs #4080. Makes dojo.declare more extensible
      remove some spruious function calls in the fast path on IE. Ensure that form disambiguation code still in play on IE and opera (where it's still, apparently, a problem). Refs #3244
      size reductions for NodeList. Refs #3961
      clobber cruft function. Refs #3961
      no one's using "makeCalled()". Removing it. Refs #3961
      size optimizations for loader. Refs #3961
      make sure that the doc system picks up _getText but that we don't pay for it otherwise. Refs #3961
      some size-related changes and a bit of crusty function removal. Refs #3961
      special case the ">" query. Was a major oversight that new tests helped dig up. Refs #3121
      IE doesn't recognize "\v" as a whitespace char to be trimmed. Refs #3121
      trimming cruft from dojo.Deferred. The toString() was especially useless. Refs #3961
      eliminates the native branch for dojo.* array methods. There wasn't a perceptable or testable speed difference and it cost us in bytes. Refs #3961
      ensure that type=dojo/connect style scripts work correctly with dijit.Declaration. Fixes #4199
      make Joe's license terms 100% specific. Refs #2619
      initial cut at a fix for default table styling backoff. Adds new table-specific rule dojoTabular which is opt-in. Refs #4209
      start of checkins for syntatic sugar for dojo.NodeList. This doesn't directly add sugar, but other methods will want/need dojo.connectPublisher(). Refs #4205
      start of some syntatic sugar for NodeLists. Makes connecting to common events easier as well as changing css class names on a list. Refs #4205.
      allow dojo.forEach(), dojo.map(), and related functions to take strings as functions to execute. Need to investigate caching the resulting anonymous functions. Makes operating on NodeLists much easier. Refs #4205
      IE methods were being plucked out of the wrong scope. Fixes #4226
      NodeList's aren't instances of the parent docuemnt's Array class. Use a test that encodes this data. Refs #3121
      make sure that deferreds get results. Refs #3121
      fixing scoped execution of connections for Declarations. Refs #4231
      make the sorting look right and re-enable row actions. Refs #4217
      ensure that dojo.publish() does not stop execution of code following it if the topic doesn't exist. Fixes #4447
      add support for onmouseenter and onmouseleave to all DOM Nodes which are dojo.connect()-ed to. Fixes #4307
      add onmouseenter and onmouseleave to the shortcuts provided by NodeList. Refs #4307
      merging patch from Mike Wilcox (mike at sitepen.com) that adds object inspection and a popup mode to Dojo's integrated Firebug Lite. Excellent patch, needed only the most minor style cleanups. Great work. Fixes #4589
      make logging before the console is available not bomb things out. Refs #4589
      fixes #4336
      D.O.H. uses clearTimeout as well as setTimeout. Support both. Fixes #4621
      merging patch from jonu to allow custom errors in Deferreds. Very useful. Fixes #4337
      catch the exception in json comment filtering. Refs #3121
      re-enable onMouseEnter/onMouseLeave, but this time with some smarts about FF's bugs. Fixes #4307
      fixes #2352
      make the console usable on IE 6. Refs #4589
      make sure that exceptions and such don't fake IE out WRT base URL. Refs #4806
      make sure we're accepting string args for the dojo.*Class methods. Refs #4580
      updating docs to note string args. Refs #4580. Thanks to ptwobrussell and phiggins for staying on me about it = )
      output something saner for most object types. Fixes #4823
      Fixes #4821. Fixes #4378. Now NodeLists are just arrays augmented with extra methods. splice() and slice() are still oddballs, but will be fixed shortly (in the same manner).
      ensure that dojo.query passes back a node list if it's inadvertently passed in. Fixes #4825
      ensure that combinations return a wrapper animation. Also, move all animations to be keyed on a single timing loop for smoother animations when mutliple animations are joined (as in dojo.fx.combine()). Fixes #3477
      clobber the "gearsObj" stray global. Refs #4839
      clobbery stray global "param" name. Refs #4839
      removing a stray global in the test. Refs #4839
      usage -> examples. Refs #4820
      add spacing for Safari. Fixes #4325
      adding first extension file for dojo.NodeList. Refs #4480
      making the constructor smaller ('cause we can), adding slice/splice, and moving animations out to dojo.NodeList-ext. Refs #4480. Refs #4205. Fixes #4822
      tests for slice/splice. Splice isn't working correctly on IE 6 yet (still digging). Refs #4822. Refs #4205
      expanding documentation and reducing file size for dojo.NodeList. Refs #4820
      minor doc update. Refs #4820
      ensure that if we're in an iframe and the parent has a console that we use it instead. Refs #4823
      complete replacement for the tokenization strategy of the query engine. Moved from ad-hoc substring searching to an AST generation/consumption system. The query engine should be significantly more robust in the face of whitespace and funky formatting than before. Code size change should be minimal despite the large-scale change. Expanded tests and a first stab at inline docs for dojo.query. All unit tests pass on FF2, Safari3, and IE 6 and 7. Fixes #4365. Fixes #4640. Fixes #2423.
      re-enable query optimizer. Refs #2423
      make direct descendant selectors more robust when they occur at the end of a selector. Fixes #4364
      make sure we get animations to combine. Fixes #4924
      make the bootstrap smaller through scope protection closures now that the doctool is picking everything up. Big props to Neil for making this possible (and safe). Refs #3961
      allow colons in IDs, names, etc. Fixes #3520
      updating for removed "!=" selector in dojo.query. Also, size reductions. Refs #3961
      cleaning up a stray debugging call, removing the non-standard ":contains" selector, fixing ":empty", and adding tons more unit tests. Refs #3121.
      ugg. I keep forgetting to re-enable the query optimizer! Refs #3961
      docs for dojo.loaded and dojo.unloaded. Refs #4820
      make dojo.delegate public and add docs. Refs #4820
      add back in docs for deprecated() and experimental(). Refs #4820
      updating inline docs such that they get picked up by the parser. Refs #4820
      minor nits. Refs #3961. I'm still very unhappy about how much space the gears detection code takes up.
      keep "pStart" out of the global namespace. Refs #4839
      keep declare() from leaking stray globals when there's no superclass. Updating docs for parser conformance. Refs #4839
      go with "example" instead of "usage". Refs #4820
      clean up doc formatting and a bit of structure. Also some code size reductions. Refs #4820. Refs #3961
      fixes editor drop-downs. Refs #4171
      good spot by Neil. Clobbering stray global. Refs #4839
      minor doc updates. Refs #4820
      minor doc update. Refs #4820
      minor doc updates and formatting for debugging. Refs #4820. toJson() no longer tries to handle DOM nodes. Fixes #4186
      documentation updates. Refs #4820
      documentation updates. Refs #4820
      minor formatting updates. Refs #4820
      minor doc formatting updates for readability. Refs #4820
      clean up doc formatting. Refs #4820
      comment formatting cleanup. Refs #4820
      update for dojo.cookie() docs. Refs #4820
      doc updates to ensure that we capture the optional property bags correctly. Refs #4820
      updating Safari version detection code to give us a real major version number. Using that update to ensure that we get/set iframe properties correctly in dojo.io.iframe. Fixes #5012
      introduced an error just prior to cutting 1.0 in an effort to debug iframeio for Safari 3. Refs #5012. Fixes #5021
      clobber spurious debug calls. Refs #5057
      re-enable :contains, expose the pseudo-match list to extension through dojo.query.pseudos, and enable sibling selectors (and tests for them). Fixes #5022. !strict
      fixes #5506. When query() is called without any params, the source list should be passed back. !strict
      ensure that the position default is actually respected. Fixes #5542. Adds better docs. !strict
      expand docs for registerModulePath(). Refs #4820
      adding support for Adobe AIR. Thanks to SitePen and Adobe for these excellent patches. Tests in next commit. Refs #5723 !strict
      update Dojo's Safari detection to use WebKit tech versions if explicit Safari versions aren't available. Fixes #5575. !strict
      update dojo.style() to accept a style bag as the second param and ensure that the mapping in dojo.NodeList is smart enough to handle it. Add and update tests. Fixes #5511. !strict
      update dojo.style() inline docs with examples. Refs #5511. !strict
      update dojo.query to defer to [node].querySelectorAll() if it's available. Fixes #5832. !strict
      more accurate detection of WebKit, Safari, etc. Un-breaks the editor. Fixes #5575. !strict
      drop base size a bit. Refs #3961
      merging instantiate() patch. Refs #5603. Needs tests.
      ensure that animations which end "naturally" decrement the timer and attempt to cancel the global interval if apropos. Fixes #5662. !strict
      applying patch from Mike Knapp for a really janktastic IE issue when using MSHTML in embedded environments. Fixes #5904. !strict
      fix dojo.isString() for IE to correctly handle cases where "" is passed. Fixes #5910. !strict
      silly IE6... Fixes #5910. !strict
      updating isFunction w/ same error handling as isArray has for null being passed on IE. Refs #5910. !strict
      adding NodeList.onfocus(). Refs #4205. !strict
      cleanup docs for fx functions. Add some more examples. Refs #5962. !strict
      Updates to the animation system and a new convenience method. Fixes #5985. !strict
      lastIndexOf() was totally borked. This fixes, rolling back changes from r10203. Refs #3961. !strict
      inline doc cleanup for the default loader. Refs #5962. !strict
      shorten NodeList code, add toggleClass(), and update documentation. Refs #4205. Refs #3961. !strict
      make sources configurable through passed properties for *all* attributes. Update and clarify docs. Refs #5962. !strict
      updating inline documentation for the bootstrap. Refs #5962. !strict
      update documentation for browser detection variables. Refs #5962. !strict
      update core documentation for query, re-enable the query optimizer, update misc docs around the rest of dojo.js as well. Refs #5962. !strict
      comment cleanup and code shortening. Also expanding test cases to handle default param packing. Refs #3961. Refs #5962. Refs #3121. !strict
      clobbering crufty djConfig values, adding docs for the stuff we actually use on a regular basis, and updating DOH to give us better command-line info. Refs #5962. !strict
      adding sugar to dojo.attr(), and therefore to it's mapping into dojo.NodeList. Expanding tests and docs. Supports a property-bag style for attrs the same way we do with dojo.style(). Refs #4205. !strict
      ugg. Fingered the wrong test file when checking in [12899]. Refs #4205. !strict
      update form processing to work on IE. Fixes #6113. Updates test page and docs in dijit.form.Form as well. !strict
      removing crufty dojo.back options which are no longer necessaray now that we've changed dojo.back to use an explicit init() method. Refs #3121
      query tests probably shouldn't be using debugAtAllCosts. Refs #3121
      minor doc formatting update. Refs #5962 !strict
      merging pottedmeat's history, object printing, and recursion changes. Fixes #6166. !strict
      update query() to move combinator-rooted queries off to a non-querySelectorAll branch until the webapis WG decides that they're the right thing to do (which they clearly are). Refs #5832. !strict
      fixes dojo.style() calculation of computed heights on IE 6 when node.style.height = "auto" is set explicitly. Skip deprecated getBoxObjectFor() call for FF3 for performance and forward-compat. All unit tests run on IE 6, FF3, and Safari. Fixes #6143. !strict
      use Firebug Lite in popup mode when debugging NodeList-fx stand-alone. Refs #6143
      seems focus()/blur() isn't synchronous on IE. Adjust test accordingly to get it to pass. Fixes #4811
      clean patch for #5617 that avoids timer for all but IE 6. Fixes #5617. !strict
      make the bootstrap use built-in console.log if there is one in order to print out debugging output. Useful on Safari and Opera. Refs #6255
      make the console detection code smaller. Fixes #6255
      ensure that we don't blow up on Safari 3.1. Fixes #5832. !strict
      updating code and docs to enable the addOnLoad(obj, function(){}); style which other dojo APIs handle. Some work to reduce code size as well. Fixes #5404
      merging adam peller's patch for infinite loop catching. Fixes #5961. !strict
      update docs and default behavior for dijit.Form and dojo.attr to make it crystal clear that unified Dojo event handling is in play when using event handlers specified by dojo.attr() and that browser-specific hacks may no longer work as expected. Fixes #6280. !strict
      update to ensure that re-setting an event handler via dojo.attr() correctly handles dis-connection. misc size reductions. The case where dojo.connect() sets a handler and then dojo.attr() sets one will NOT be handled. Updates to tests. Fixes #6310. Fixes #6234. !strict
      fixes animation unit tests for Opera 9.2. Fixes #6325. !strict
      update timings to allow unit tests to pass on slower browsers (e.g., IE 6, etc.). Fixes #6330
      add more time to get the fx tests to pass on IE 6. Refs #6330
      remove spurious logging statements. Refs #6325. !strict
      roll back wipeIn changes. Don't need 'em and they were borking the dijit test pages. Refs #6325. !strict
      merging Sam Foster's html.set() patch + tests. Refs #5647. We can close when ContentPane is refactored to use html.set()
      merging great patch from Liu Cougar for correctly handling the "3n-3" style of CSS selectors and fixing my broken logic on earlier selectors like "0n+1". Fixes #6418. !strict
      merging great Firebug Lite enhancements patch to add DOM inspection, kill memory leaks for good, and lay out objects more teresely. From Mike Wilcox of SitePen (CCLA on file). Fixes #6619. !strict
      fixing mixing variable. Refs #6619. !strict
      re-add a close button to Firebug Lite. Fixes #6680. !strict
      updating the overall font size to be 12px as discussed in the weekly full-team meeting. Fixes #6755
      updates for IE 6. Refs #6755
      adding non-destructive filters updating for opacity setting. Also ensuring that we tickle the zoom bit. Also adding tests to ensure that we handle dojo.attr(node, "class", "..."); correctly. Fixes #6618. Fixes #6957. Fixes #6937. !strict
      removing crufty zoom setting code which is now rolled into dojo.style() for IE. Refs #6937. !strict
      updating inline docs, adding examples, etc. Refs #5962. !strict (like most things in _base)
      update docs and unit tests. Also ensure that scoping is correct in dojo.string.substitute() formatters. Fixes #6247. Refs #5962
      clobbering duplicate code in JSON encoding block. Refs #3961 !strict
      getting even more miserly w/ the bytes. Refs #3961 !strict
      some more byte scrimping. Refs #3961 !strict
      size saving patch from phiggins. Nice work. Refs #3961 !strict
      merging performance patch from Tom to speed up _toArray calculations (which many NodeList operations spend much time in). Fixes #6722. !strict
      add case-sensitive searching to dojo.query() for XML document. Tested on IE, FF, and Safari. Fixes #3866. Fixes #5262. !strict
      adding support for the ":checked" pseudo selector. Fixes #5179. !strict
      ensuring that we search cssText instead of the style property directly when doing string matches on attribute values. Fixes #7037. Tested on IE, FF, Safari, and Opera 9.5. !strict
      ensure that cssText works as intended. Fixes #2855. !strict
      ensure that cross-document xpath-run queries don't bomb out on FF3 due to new security restrictions. Fixes #7075. !strict
      fixing _toArray() for DOM collections on IE. Fixes #7178. !strict
      fixes #7205. Ensures that we detect case sensitivity correctly on XML documents in FF and Safari when they're passed in directly as the root. !strict
      adopting Cornford optimization for dojo.delegate. Thanks to Neil for the patch. Fixes #7282. !strict
      update dojo.attr() to handle the case of innerHTML and style: {...} being passed in as values. Fixes #7305. !strict
      make sure that dojo.style() returns "bold" and not 700 for font weight queries on IE. Refs #7480. !strict
      remove spurious debug(). Refs #7480. !strict
      dev-time support for detecting and loading a XUL-based host-environment. Also adds a host environment file build specifically for use in extensions and a trivial patch to allow the build system to be told that it's in a browser-like environment.
      major addition of comments to the dojo.query() code as well as some small cleanups to enable query portability. Adding a test page for using query() without the rest of Dojo. Fixes #7794. !strict
      make the default loader work in strict XHTML environments and fix Opera to correctly detect case sensitivity in XML/XHTML environments for queries. A small PHP script is added to test these. Sadly, both Safari and Opera don't observe script ordering for injected tags, so we had to make the dev loader script a bit more complicated to accomidate. Fixes #7455. Fixes #7214. !strict
      ensure that attribute names and values are correctly looked up in XML documents on IE. Will need to merge this back into the 1.2 branch. Refs #7796. !strict
      build system fix to prevent us looking for nodes in environments that don't have them. Refs #7805. !strict
      fix to prevent Firebug Lite from popping up in Firefox Minefield builds. Fixes #7818. !strict
      removing bad impls of createElement() and elem(). Refs #7802. Refs #3833. !strict
      making sure that opacity animations work correctly in XUL environments. Fixes #7942. !strict
      ensure that firebug lite doesn't throw errors when logging text nodes. Fixes #8008. !strict
      adding dojo.pushContext and dojo.popContext calls for use in the XUL host environment.
      removing the QSA branch. Fixes #8102. !strict
      ensure that if nothing is passed to mixin() that a blank object is still returned. Fixes #8335 and provides tests.
      The new query engine (aka: "steak").
      some require() and NodeList doc changes that happened while I was hacking on steak. Refs #7072, #4425
      steak -> acme. Need a better name. Refs #7072. Refs #4425. !strict
      updating for the acme engine. Refs #7072. Refs #4425
      fixing regression in XML tests related to tag name case sensitivity. Fixes #8356. !strict
      some size reductions. Also (finally) enables the QSA branch. Not sure how I missed that.
      ifdef's for the core to create a webkit-only version for use in, e.g., mobile devices. Refs #8447
      fix test html unit test failure in webkitMobile profile. Refs #8447. !strict
      my bad. Refs #8447. !strict
      tighten it up just a bit more for the webkitMobile build. Refs #8447. !strict
      more mobileWebkit exclusions. Refs #8447. !strict
      remove the internal namespace protection closures since we get a global one generated for us by the build system. Also, remove the shortened names for Dojo where possible so that shrinksafe can alias "dojo" on its own. Saves us another couple hundred bytes or so after gzip and should speed eval time.
      fixes build breakage that haysmark and phiggins found. Refs #8447. !strict
      fix for IE in non-build environment. !strict. Refs #8447
      handle the case where a failed scope lookup would otherwise cause us to search the entire docment. Instead, return an empty NodeList. Adds test for this case. Fixes #7211. !strict
      hrm, not sure how this got missed... Refs #8447. !strict
      change the behavior of addOnUnload and addOnWindowUnload to prevent default registration of unload and onbeforeunload handlers since they can defeat fastback behavior. Updating docs for affected APIs and adding a test of the system to verify that it works. The test is kind of a pain to run, but it does proove that the change has the intended behavior.
      removes cruft, updates code comments, and makes some operations slightly quicker. Refs #7072 !strict
      make sue that query-sizzle.js can still run stand-alone (for side-by-side perf tests). Refs #8547 !strict
      exposing _mapIntoDojo, but renaming to _mapIn. Fixes #7295. !strict
      adding existance test to ID filtering. Fixes #7368. !strict
      test case for #7125. Refs #7125
      make sure that null lists don't fool us in the DOM branch + test. Fixes #7125. !strict
      make sure that create() works with actual markup. Fixes #8613. !strict
      adding sugar to dojo.NodeList instances. Refs #8612. Needs expanded tests. !strict
      fixes a typo in getNodeIndex which caused node indexes to be lost, dropping our speed on nth-child queries. Also significantly improves peformance on ":not" pseudos by specifying ignores in the test generator. Fixes #8654. !strict
      even more speedups on the DOM branch. Was doing some dumb stuff for ID and tag-only queries which this checkin fixes. Also, attempt to preserve zip-killing across groups. Tests show us to be ~30% faster than sizzle on IE 6. QSA branch still faster nearly everywhere, but by variable amounts depending on the engine (15-50%). Refs #8654. !strict
      test case. Refs #8775
      correctness fixes in the DOM branch. Make sure we don't double-count things in "," separated queries and ensure that we don't use QSA in suspect documents on WebKit (pending webkit fix to case-sensitivity issue). Fixes #8775. !strict
      fixes Prototype's bug for document.getElementsByClassName(), opens up our margin on byId-rooted tests on QSA, updates the case-sensitivity bug tests, and brings the docs into alignment w/ the code. Fixes #8697. !strict
      nth-child selectors in the DOM branch were potentially returning the wrong results since the simpleElementTest would be elided for firstChild nodes that weren't elements. This is fixed by looking for firstElementChild on browsers that can handle it. Fixes #8775. !strict
      removing extraneous parseInt calls. Refs #8775. !strict
      re-enable index cache. Refs #8775. !strict
      style cleanup, doc expansion, fixes minor nits. Refs #8606. !strict
      fixing html.js formatting. Refs #9063. !strict
      doc comment formatting cleanup. Refs #9063
      create (and use) a local reference to dojo.byId() in order to boost addclass-odd test #'s on IE6. Very strange, but the performance difference is nearly order-of-magnitude (80ms vs > 800ms).
      fix for query system matching too many elements (possibly not checking for uniqueness in a set). Refs #9071. Back-port to 1.3.x branch will fix. !strict
      Inline isString checks in dojo.js to boost performance.
      Adds a unit test to ensure that we get intrinsic array methods passed through fine and adds a check in the wrapping method to ensure that we're always wrapping real Arrays. A behavior change in FF 3.5 seems to be causing the assignment of constructors to objects to keep prototypes from being inherited along with.
      missed a spurious change in the last CL. Refs #9499
      Refs #8775. I expect this to break again.
      comment fixes I had laying around. !strict
      expanding the docs for `dojo.connect` to be understandable by humans. !strict
      more doc updates. Refs #8606. !strict

Becky Gibson (5):
      refs #6665 Updated _fixAttrName and hasAttr to deal with idiosyncracies of htmlFor and for attributes.  Created tests. !strict
      fixes #8557 #8559  replace the icons that only use color with new icons that use color and different shapes.  Also provide text only equivalents for when system is in high contrast mode.
      fixes #9170 when column reordering in grid and high contrast mode is detected, use the default DnD icons which are low vision accessible. Updated default dnd to add generateText parameter so can turn off generation of text node in the icon. !strict
      refs #8991 - added tests for disabled attr on input element since referencing disabled via property or via getAttribute("disabled") returns different results on different browsers. This test would catch any changes to dojo.attr() to not reference disabled as node.disabled.
      refs #8991 updated disabled test to use valid markup (disabled="true" is NOT valid, disabled="disabled" is valid

Ben Hockey (1):
      removed an unnecessary calculation of _endTime and removed an unused variable _globalTimerList.  removed some trailing whitespace.  passes unit tests and fixes #11218 !strict

Ben Lowery (20):
      fixes #3059. add in proper provide
      switch to _base class functions. fixes #3087
      fixes #3133. force extractRgb to return an array of number instead of an array of string for rgb(0,0,0) style strings
      refs #3133. remove unneeded function wrapper
      fixes #3742. add support for native array iteration methods for safari3
      refs #3742. fixup little whitespace issue.
      refs #3769. add a test case that allows array to be tested from a browser-based environment.
      refs #3769, refs #3742. Revert change to enable native array iteration methods
      fixes #4586
      testcase for delay. refs #4586
      Fixes #4741. Reduce timer to 200ms, matching 0.4's impl
      Fixes #4647. Comment out calls to console, except for diagnostic call for x-domain.
      Fixes #4777. Consistent hash handling and test cases added to dojo.back
      refs #2914. Fixes initial state storage by using normalized getHash function, scopes safari2 history hack, speeds up checkLocation just a tick
      fix up docs a bit. refs #2914.
      they're not wrong, they're just in need of some love. refs #2914.
      fixes #6614. makes trims that result in one character and have trailing whitespace work for dojo.string.trim
      Fixed up a doc bug. fixes #9189. !strict
      refs #9942. Adds in doc updates for the other array iteration methods indicating our difference with sparse arrays.
      refs #9942. Minor doc consistency fixes for the sparse array warning.

Bill Keese (104):
      Fix typo in previous checkin to this file.
      fix typo.  refs #2500
      Unit test for dojo.map.  Not working on IE.  Refs #3243.
      Fixes #3243, which was failing because dojo.map() wasn't passing the idx and this
      Refactor dojo.dnd._getOffset(), fixing problem with border calculation and
      More dojo.query() tests.
      Every node on IE defines "removeChild" as a native attribute,
      Declare url as a string for the benefit of the parser.
      Standardize on "label" (not caption).  Fixes #3288.
      dojo._destroyElement argument doesn't necessarily have a parent node,
      Add tests for parsing "checked" and "disabled" attributes.  Refs #3633.
      Make slideIn()/slideOut() assume a natural height for the node, rather
      Fix various IE problems with parameters that are builtin function names,
      Following dojo (core)'s lead, make dijit.js include certain modules by default,
      Fix comments (evt is obviously not optional).  Refs #2556.
      Fix *re*-execution of animations with start/end points specified as functions.  The function references were getting overwritten on the first execution.
      dojo.style(node, "height") can return "auto" or "" on IE rather than a number,
      Fix comment to match code.  Fixes #3869.
      Fix parser bugs reading disabled/checked, and also b="false" and foo="" type
      On IE7 Alpha(Filter opacity=100) makes text look fuzzy so remove it altogether.
      Fix setSelectable() on IE to affect the selectability of all the subnodes.
      make sure that map returns a NodeList.  refs #3244.
      Fixes #4338: allow setting widget startup methods via <script type="dojo/method" event="foo">
      Disable onmouseenter and onmouseout until #4307 is fixed on FF.  (Refs #4307.)
      Fixes #4325 on trunk: typo in safari version number about messing up dojo.abs() and thus drop down/popup placement
      Fix regex for removing opacity FILTER() clause when opacity==100%.  Fixes #5371, refs #2661.
      Fix dojo.coords() on FF to account for borders  (rather than returning the content box).  Fixes #3222, #3676, #5541.  !strict
      Modify API doc of dojo._abs() to match behavior on FF.
      Fixes #5857, #5860: dojo.coords() wasn't accounting for borders an ancestor nodes (safari bug only) !strict
      Refs #5858: _base/test_FocusManager.html "restore focus" button failing.
      Remove connect.html, which is testing something that isn't supported.
      <link ... /> is invalid, should be <link ... > (unless you are writing XHTML).
      <link ... /> is invalid, should be <link ... > (unless you are writing XHTML).
      Fixes for cross-window widget creation.
      Fixes #6928: fix parser so it handles empty arrays.
      Fixes #6940: tree items cutoff at right for tree wrapped in narrow div
      Fix wipeOut() to also leave the overflow setting as it was before the effect started.
      Supplementary tests for dojo._abs().
      Assume that value returned from script IO is UTF-8.
      Fixes to dojo._abs().  Working much better than before (works
      Further patches to dojo._abs() for opera and FF2.
      Fixes #6957 (!strict): dojo.attr() does not handle "class" correctly on IE.
      Fixes #6772: dojo.hasAttr busts when node doesn't have the getAttributeNode function
      Add workaround to long-standing FF2 drag problem.
      Fix comment; refs #6755.
      Fixes #7748, refs #6876: reinstate faux onmouseenter / onmouseleave events for safari.
      Fixes for dojo.attr(node, "class", ...) on IE8, plus adding some hints to test file.
      Don't call _getIeDocumentElementOffset() on IE8.  This fixes the tests/_base/abs.html test.  Also various marginal size reductions.
      Both opera and safari support getBoundingClientRect(), so remove special code for safari and opera in the else branch (which computes results manually by tracing the DOM tree).
      Oops, safari doesn't support getBoundingClientRect() after all.
      Fix event normalization (for mouse position) for IE8.
      Call _getIeDocumentElementOffset() on IE8 again, but it's returning (0,0) until we can make it return the correct value.  The IE7 branch of that function doesn't work on IE8 (in strict mode) because it's result is affected by margin on <body>, and negatively affected by body's scroll.
      Make getIeDocumentElementOffset() return correct results for IE8.
      Fixes #8247: dojo._abs returns wrong x, y in FF 3 when there is an html border.
      dojo._abs was returning wrong x, y in Safari (v3.2.1 on the mac) when there was an html border...
      Fix dojo.coords() on IE in quirksmode.   Patch from Nic (CLA on file), thanks Nic!
      IE8: Fix problem with dojo._abs() when page is scrolled.
      Fix doc typo, refs #8628 !strict
      Test infrastructure updates, as discussed with Mark:
      Only do try/catch block for relatedTarget.tagName for FF2.   It's not needed for FF3 and causes some issues.
      Parser was putting "constructor" and "toString" into the params passed to every widget.
      Testcase that parser doesn't pass unwanted parameters to a widget.
      Roll back line shortening changes from [17195] (refs #9063); leaving in lint fixes and changes to API comment formatting/content.  !strict as usual.
      Test that null/undefined arg to dojo.byId() returns a null value.
      Use spaces rather than tabs before keywords as discussed on dojo-contributors.
      Comment formatting as per dojo style guidelines, refs #7390 !strict.
      Fix comment formatting as per dojo's style guidelines, and to make the doc parser work.
      When dragging between two trees with the same model, move the item rather than copying it.   Fixes #7651.
      Automated test for keyboard event handling, showing safari 4 problems with ESC and ctrl-q.   Refs #9511, #9527, #9528.
      Fix dojo.coords() miscalculation on FF3.5 due to page margin, fixes #9602 !strict.
      Update DnD API doc, in particular about the ({id: 123, data: "hello world", type: ["text"]}) type objects that are passed to / returned from various methods.  I've named this dojo.dnd.Item since it's returned by getItem(), although it shouldn't be confused with dojo.data.Item.  (refs #9684).
      Fix API comment, dojo.dnd.Item really just has "type" and "data.  Refs #9684 !strict.
      Widgets like BorderContainer add properties to _Widget via dojo.extend().  If BorderContainer is loaded after _Widget's parameter list has been cached, we need to refresh that cache (for _Widget and all widgets that extend _Widget).
      Convert dijit._Calendar references to dijit.Calendar.
      Add comment that dojo.isArray() doesn't work across page boundaries (on IE).  Refs #5334 !strict.
      Make dojo.byId(node) on IE work for nodes from other documents again; it stopped working in [17266] (refs #9114).   Using typeof check rather than dojo.isString() for performance (refs #9131).
      Test was invalid since it ran doh.is() asynchronously without a try/catch block, causing an unhandled exceptionrather than calling d.errBack().  Was popping up error dialog on IE and hanging the regression.
      Remove vestigal comment, refs #9862 !strict.
      Seems there's no need to create a function in the global scope everytime there's code in markup like onclick="console.log('hello world')"... creating a local variable works fine and avoids the IE memory leak.  Refs #10206 !strict.
      One more leak fix for firebug lite, refs #10208 !strict.
      Move dijit sniff code to core, as dojo.uacss.  As before, doing a dojo.require("dojo.uacss") will add the relevant classes to <html>.
      Move dijit's getViewport(), getDocumentWindow() and scrollIntoView() functions into new dojo.window core module.
      Some name changes:
      Now that the needed methods are available in core, move DOH robot's _position() and _scrollIntoView() methods from dijit/robot.js to dojo/robot.js.
      Remove parser dependence on dojo.query() by implementing search for dojoType and <script type="dojo/..."> using a simple recursive algorithm.
      Enhance parser to track dir=ltr/rtl settings on nodes (with or without a dojoType setting), and propogate those settings down to descendant widgets when the widgets are constructed.
      Add "inherited" parameter to parser to specify default dir/lang for created widgets.
      Since we no longer cache the value in _parentWindow (see Cougar's comment about avoid a memory), there's no reason to check if the value is check that variable.
      just fixing doc typo, the first parameter to watch() is optional, not the second, refs #10839.
      just fixing doc typo, the first parameter to watch() is optional, not the second, refs #10839.
      If widget prototype has a "stopParser" flag set, then don't look for widgets declared inside of object.   Still looks for <script type=dojo/*> though.   Refs #10402 !strict.
      [21680] assumed that all classes are defined before the parser scans the document for nodes with dojoType set.   That isn't the case for classes defined via dojo.Declaration.
      When parser called on root of document, be sure to get the dir setting of the document itself as the initial inherited dir value.   This fixes the order of the icons in test_Toolbar.html?dir=rtl.   Refs #10402 !strict.
      API doc for hash based sets, refs #10839
      Test changes corresponding to [21748], which passes in dir all the time.  Refs #10402.
      fix backwards compatibility issue for dojo.parser.instantiate() on nodes that specify type (like <input type=...>), refs #10402, fixes #10951 !strict.
      fix corner cases of onClick="return" and onClick="", patch from Doug, fixes #11005 !strict.
      Add flag for quirks mode for non-IE browsers.  Both FF and safari change behavior slightly in quirks mode: on an <input> with CSS of border: 1px, width: 16px, in quirks mode the width includes the border (and the content-box width is 14px).  In general FF/safari still do content-box sizing, but not in that case.   Refs #7028 !strict.
      Various test fixes:
      Apparently on FF, setTimeout(foo, 500) may fire slightly before 500ms, similar to how sleep(10) in C can fire any time after 9s.   Fixes intermittent FF test failure by ignoring slightly faster execution than expected.   Fixes #11402.
      Naked DOH assertions in asynchronous code are lost.  Need to surround them with d.addTestCallback()/d.addTestErrback() (or a custom try/catch calling d.errback() on failure).  Refs #11402.
      patch from Adam to ignore timezone, fixes #11198
      Fix parser to obey _noScript flag in prototype again, so that dijit.Declaration works.   Also adding unit tests for dijit.Declaration.  Fixes #11586 !strict.
      IE10 even in quirks mode does not support filters for opacity.  Fixes #16117 !strict.  Backport to 1.5.

Brad Neuberg (2):
      fixes #3339: dojo.isGears incorrectly initialized
      fixes #3440

Brian Skinner (11):
      JsonItemStore: replaced some spaces with tabs, and removed attribute-item in doc strings. refs #2891, #3291
      Fixes #3369
      fixes #3627
      fixes #3629
      refs #3632: [dojo.data] in JsonItemStore file format, change {reference: to
      refs #3630
      fixes #3654
      fixes #3669
      fixes #3669
      this time for sure...
      fixes #3873 -- Bug in ItemFileWriteStore that would show up when using a file that did not specify an identifier and then calling newItem() after calling deleteItem().  Changed deleteItem() to place null values in the _arrayOfAllItems, rather than removing elements from the array.  Should also improve performance of deleteItem() and revert().

Bryan Forbes (17):
      refs #2710
      refs #2710
      refs #2609
      refs #2710
      refs #2710
      refs #2710
      fixes #4239 - code size reduction and optimization for dojo.fx.slideTo.
      fixes #5276. fixes #5315.
      fixes #5335.
      Added tests for combine and chain.  refs #3121.
      fixes #5884.
      refs #5884.
      fixes #5634. !strict
      Set svn:eol-style to 'native' on files missing it. refs #7776 !strict
      refs #8171
      Set svn:eol-style to 'native' on files missing it. (refs #7776 !strict)
      Set svn:eol-style to 'native' on files missing it (refs #7776 !strict).

Chris Barber (4):
      Added support for Appcelerator Titanium to the loader. Fixes #9405.
      Added whitespace to _isDocumentOk(), refs #7390.
      Added inline documentation for the djConfig.useCustomLogger.  Refs #8606.
      Added a flag for which console methods are stubs created by the bootstrap. fixes #9907 !strict

Colin Snover (2):
      Use per element determination of children vs childNodes, fixes #12488 !strict
      Updating metadata for 1.5.3

Douglas Hays (42):
      Fixes #3378.  Change clientWidth/Height to offsetWidth/Height in _getContentBox and adjust calculations accordingly.
      References #3621.  Change array delimiter in the parser to comma from semi-colon to align with JSON delimiters and to allow HTML special characters like  
      References #3621.  Fixed parser test page to use commas instead of semi-colons.
      Fixes #2982.  Changed ele.id to ele.attributes.id.value.  Added a testcase for the IE getElementById bug.
      References #3985.  IE doesn't always set item.selected=true for the value attribute.
      References #4034.  Change the str2obj to return NaN for value="" wrt number attributes.
      References #5607.  Added function attribute support to dojo.attr. !strict
      Fixes #5620 !strict.  Added support to dojo.attr for boolean attributes on Firefox.
      Fixes #5631 !strict.  Add explicit test for boolean and function types in dojo.attr.
      Fixes #5781 !strict.  Boolean attributes need to use the javascript value instead of the DOM attribute in dojo.attr.
      References #5562.  Add timer to Moveable to reduce calls to dojo.marginBox to help IE performance.
      References #6245.  dojo manual testing front-end.
      References #6667 !strict.  Add new charOrCode member to key event object to simplify code, removing the need to check both charCode/keyChar AND keyCode when looking for specific keys.  Reviewed by sjmiles.
      References #6879 !strict.  Fixed bad DOCTYPE.
      Fixes #7216 !strict.  Check for a valid node before calling dojo.isDescendant to avoid FF2 bug.
      References #3486.  Proxy commit for haysmark.
      References #7391.  Add robot.css import to test file.
      References #7398.  Remove xhr call to check for applet existence.
      References #7441.  Proxy commit for haysmark.
      References #7521. Proxy commit for haysmark.
      Fixes #6876 !strict.  Force safari to take a similar code path as IE and generate faux keypress events.  Removed the mac keycodes since safari 3.x uses the same keycodes as IE and FF.
      References #7681.  The robot code should execute with the same BASE as the page being tested so that XHR gets work correctly (Safari 3.1 seems to require this).
      Fixes #8344 !strict.  Change readonly to readOnly for all browsers.
      Fixes #8367.  Per jburke, unset cached dojo._bodyLtr in withGlobal and withDoc.  Had to call delete since _isBodyLtr returns whatever value is set (null/undefined/true/false).
      References #6245.  Remove manualTests from dojo/ in favor of creating robot tests.
      Fixes #8776.  Resize the robot's test page iframe and console windows such that the container browser window does not need its own scrollbar.
      Fixes #9596 !strict.  Add check in _isButtonTag for null returned from getAttribute("type") since INPUT tags can default to type="text" when not explicitly specified.
      Fixes #5197 !strict.  Change _isBodyLtr to inspect DOM nodes instead of computed style.  Add html_isBodyLtr.html automated tests.
      Fixes #9617 !strict.  Optimize _docScroll to reduce needless evaluations.  Add html_docScroll.html automated test.
      Fixes #9648 !strict.  Add dojo.position (replacing _abs) to core that returns the x/y start of the node's border-box, and also w/h for the border-to-border size.  Most uses of coords() should be changed by component/file owners to use either position() for border-box calculations or marginBox() if the margin-box is needed.  The mixed box type of coords is probably a bad programming model for users to adopt.  Also added some automated tests for the new position method.  I'll update t [...]
      References #9648.  Add new position() to NodeList processing along with DOH testcase.
      References #9648.  Add new position() pseudo-doc to NodeList.
      Fixes #6184, #9528, #6184 !strict.  Forced Esc key to generate fake keypress on WebKit.  Removed keydown special WebKit/Esc processing from dijit's popup.js and _FormWidget.js.  Verified that dijit's robot tests ComboBox_a11y.html, Dialog_a11y.html, and Menu_a11y.html now run without errors on Safari4/Windows.
      Fixes #10020.  References #9859.  Fixed several problems related to dojo.position()'s use inside iframes.  Fixed several IE-specific problems related to inaccurate position() attributes caused by iframe style.  Fixed withDoc to restore _isBodyLtr and to properly set/restore isQuirks.  Changed isQuirks definition to only check for "BackCompat" since other values are for ancient unsupported browsers.  Added automated iframe tests to html.html.  Removed workaround from dijit's scroll.j [...]
      Fixes #10015 !strict.  Proxy commit for haysmark (CCLA on file).  Add multi-iframe scrollIntoView into robot API for dijit.  Add editor test that makes use of the new API.
      Fixes #9652.  Tweak test to better work with IE.
      Fixes #10095 !strict.  Change destroy() to remember the current _destroyContainer ownerDocument to keep IE from throwing an exception when accessing the ownerDocument attribute later.  Added automated testcase to html.html.
      Fixes #10147.  Add Deferred processing to html_rtl.html to give the browser time to render.
      Fixes #10312 !strict. Reorder code to call _marktargetanchor after startdrag without having to wait for further mouse movement.  Add automated test to DateGgrid_mouse.html since it was already setup to move columns and was adversely affected by this issue.
      Fixes #10971.  Reset horizontal scrolling before hiding scrollbar so that IE8 isn't confused.
      Fixes #11247.  Simplify dojoTabular CSS.
      Fixes #11575 !strict.  Change dojo.version to 1.5.1 in 1.5 branch.

Dustin Machi (32):
      refs #2915 . Reorg of rpc
      refs #2915 . Reorg of rpc
      refs #2915 . Reorg of rpc
      refs #2915. Reoorg of rpc
      refs #2915 . Reorg of rpc
      refs #2915 . Reorg of rpc
      refs #2915 . Reorg of rpc
      port rpc to 0.9 and fix some other tickets along the way, fixes #2750 , fixes #2303 , fixes #2185 , refs #2915.  There is still a json serialization bug, but its almost done
      fixup a couple of test errors, rpc tests now pass, refs #2915
      more work on rpc service. Add string arg back in for automatic retrieval of smd if desired.  moved constructor from JsonService to RpcService in preparation for enhancement. add new tests for this functionality. refs #2915
      port YahooService to 0.9 as a generic JsonPService, refs #2915
      remove trailing comma. Fixes #3322
      Rename JsonPService.js to JsonpService.js for consistency with the io system. fixes #3237
      update json serializer to produce yaml compliant output. also fixed dependent xhr test. fixes #1889
      fix some Rpc bugs. references #3111. references #2915
      fixes to some rpc tests and show a useful error for tests that we know are going to fail when accessed from file://. fixes #3111
      change the error thrown by a deferred on cancel so we can identify it is of this type. refs #3442
      fixes empyt param case in jsonpservice. fixes #3443
      add in smd to the rpc obj, fixes #3457
      add timeout param to jsonpservice. fixes #3538
      fixes #3403
      port DeferredList from 0.4 to 0.9. refs #3499
      fix scope issue with deferredLists.  fixes #4687
      fixes deferred list return problem. fixes #4240  execept issue 3 which is the expected behavior though admittedly can be confusing at times.
      add inline docs for DeferredList. fixes #3499
      update documentation and fix a few minor bugs like comments in the yahoo smd causing the tests to fail. refs #4750
      fix rpc so it can accept (convert) a dojo._Url object when passed this as its parameter.  Cleaner solution to the last checkin's little test hack.  Fixes #4750
      test commit refs #5067
      test commit refs #5067
      trac svn hooks fixed.  fixes #5067
      Moves smd v1 yahoo from dojox.rpc to dojo.tests.resources for its continued use as the test smd for JsonpService. fixes #6020
      fix deferredList. fixes #5899 and adds test file. thanks tvachon

Dylan Schiemann (23):
      refs #2423, add test showing dojo.query whitespace pickiness
      add an additional dojo.query unit test, refs #3427
      fixes #2237, handle case for RPCService where an error object should be returned instead of a string
      fixes #5399, firebug lite support to catch xml documents
      fixes #5569, NodeList.connect documentation wrong
      refs #6653, fix an accidental global references in dojo.html, and remove duplicate declaration, \!strict
      refs #6653, fix an accidental global references in fx, \!strict
      refs #6653, fix an accidental global references in dojo.cookie, \!strict
      refs #6653, fix an accidental global reference in dojo.dnd.Avatar, \!strict
      refs #6653, fix an accidental global reference in dojo.data.sorter
      refs #6653, fix an accidental global reference in firebug lite, and remove a redeclaration, \!strict
      refs #6653, fix an accidental global reference in firebug lite, and remove a redeclaration, take 2, \!strict
      refs #6653, fix a minor style guideline variation in firebug lite, \!strict
      refs #6653, fix an accidental global reference in dojo.json
      refs #6653, fix an accidental global reference in dojo.declare._extend, \!strict
      refs #6793, add information about what an error case is for the error handler for xhr
      fixes #6919, performance optimizations in array loop code, thanks schallm
      fixes #6393, prevent rhino from overriding any existing [set|clear]Timeout implementations, thanks Jordi Albornoz Mulligan for the patch, \!strict
      refs #7121, style guide cleanup for hostenv_rhino, \!strict
      fixes #7330, dojo.NodeList doesn't support .onsubmit
      add load and error syntactic sugar for NodeList, fixes #7330
      refs #8798, bundleMap, not bundle, \!strict
      fixes #10378, replace ></meta> with />

Eugene Lazutkin (256):
      ported DnD2 to 0.9 as DnD with all tests, there is a small bug related to dojo.connect(), which will be fixed later; refs #2728.
      Fixes #3116 --- undefined variable.
      Multiple style bugs prevented dojo.hex2rgb() from workin on IE. Fixes #3119.
      Making sure that dojo.blendColor() works under IE, with wider range of arguments; added all 16 HTML4 colors from the CSS3 color module --- we need to support at least those. Fixes #3123.
      Updates the alpha value of the "_cache" member. Fixes #3128.
      Replaces direct access to className to indirect getAttribute(), which is better supported. Minor performance improvement for dojo.addClass(). Fixes #3225.
      Replaces the last instance of direct access to className to indirect getAttribute(). Fixes #3225.
      Reverting the previous change. Refs #3227.
      Refs #3225. Mostly reverting previous changes.
      Minor stylistic changes. Fixes #3227.
      Bug: moving items without a current item raises an exception. Fixes #3246.
      Bug: an empty drag is possible. Fixes #3230.
      Moves the dragged node to the body to avoid esoteric settings on nodes parents. The fix works on FF and Opera, but has some subtle shift on IE. Fixes #3178.
      Switched from manual code to dojo.marginBox() for all element movements reducing the code size in the process. Thanks Scott J. Miles! Fixes #3249. Fixes #3255.
      Reverting the change to fix #3178. Refs #3178.
      Making sure that the correct document is used for move operations. Tahnk you, doughays! Fixes #3298.
      Implements enhanced styles of drag detection, and customizable Mover. Includes box-constrained, and parent-constrained movers. A grid mover is implemented as an example of custom mover. Refs #3223. Refs #3224.
      Fixing typo (thx Anni Hienola), and slightly restructuring the code to accomodate for new dojo.unsubscribe() syntax. Fixes #3331.
      For the lack of better place I ported flickr_demo.html to dojo/tests/dnd/. Fixes #3334.
      Added a missing mouse button state change. Fixes #3316.
      Increased the distance between the cursor and the avatar. Fixes #3341.
      Now ESCAPE cancels the drag. Fixes #2214.
      Added docs, changed the API for dojo.dnd2.parentConstrainedMover to make it more flexible, fixed static positioning of a parent, broken absolute positioning of a parent. Refs #3224.
      Typo fix. Refs #3297.
      More typo fixes. Refs #3297.
      Updates to accomodate new refactored html.js, and minor typo fixes. Refs #3224.
      IE-specific fixes. Fixes #3224.
      Added the customization for copyOnly sources. Refs #3229.
      Added global and local classes, so programmers can change UI during the move. Added topics to indicated that the move has started and ended. Fixes #3371.
      General cleanup and a typo fix. Thx, Scott. Fixes #3229.
      Cleanup. Refs #3229.
      Implemented autoscroll. Tested on FF, IE, Op, and Safari -- the last two suffer from giblets --- probably need to file bugs to them. Implemented a condenced version of getViewport(). It is quite small --- probably it should go to html.js.
      In order to stop critics I've implemented dojo.dnd.Target for them. Enjoy! Fixes #3568.
      Forgot to add one more thing. Refs #3568.
      Added a try guard against some FF-related quirks. See the ticket for details. Added a test for this kind of problem. Fixes #3370.
      Refactored autoscroll, minor cleanup. Implemented dojo.dnd.autoScrollNodes(), but at the moment it works only on FF --- apparently dojo.html methods do not isolate us good enough from idiosyncrasies of different browsers. Will investigate more. For the new method is not used anywhere. Refs #1597.
      Added xhrPut() and xhrDelete() tests. Corrected an IE-specific bug (the trailing comma). Refs #3121.
      Added rawXhrPost() test. Refs #3121.
      Updated TODO. Refs #3121.
      Added testClassFunctions(). Refs #3121.
      Implemented an inexpensive space trimming for class strings. Refs #3121. Fixes #3604.
      Added more tests for semi-public box functions. Refs #3121.
      Added a todo item: need tests for dojo.place(). Refs #3121.
      All topic names are normalized from the camel case to the folder case. Fixes #3649.
      Removed selectors and filters, switched to dojo.query() selection, implemented the Dojo markup for DnD (no move yet), added examples. Refs #3650.
      Minor normalization of user-supplied functions. Refs #3650.
      DnD move is adapted to be used with the Dojo markup. Fixes #3650.
      After multiple consultations I am committing a table of all named CSS3 colors. It populates (replaces) dojo.Color.named table. Because it is out of the core we don't sweat too much about its size => it is not manually compressed. Fixes #3188.
      Implementing moving and cloning nodes, when there is no creator function. Fixes #3664.
      Added dojo.trim() (one-liner), and dojo.string.trim(). Thx Steven Levithan. Fixes #3731.
      New implementation of dojo.Color as an object, support for common CSS formats, updates to utility functions, optimizations for fx.js, expanded test cases, minor updates to dojox.gfx. Refs #3652.
      Added one more test for dojo.Color. Fixed dojo.data test, which used dojo.Color. Refs #3652.
      Added full CSS3 component recognition (rgb, rgba, hsl, hsla, and rgb/rgba with percentage values).
      Reduced size of dojo.Color, removed almost all type checks, changed the signature of dojo.blendColors(), and updated dojo.fx accordingly, dojo.Color.sanitize() moved to the core, added dojo.colors.makeGrey(), changed tests, added tests for dojo.blendColors(). Thanks, Adam. Fixes #3653.
      Replacing the dndItem class with already defined and already in use the dojoDndItem class. Fixes #3846.
      Changed zIndex of avatar to 1999 --- it ensures that it is above any pop-up widgets. Minor typo fix in docs. Fixes #654.
      Minor cleanup. Refs #3875.
      Abstracted access to the data map. Fixes #3923.
      Added a parameter to handle form elements. Fixes #4032.
      Reverting back changes, which were lost during the dojo.declare cleanup. Refs #4069.
      dojo.removeClass() changes the className only if it is different after removal. Refs #4125.
      Yes, it was the Firebug-Lite all along, who triggered some bizarre bugs in IE. This simple fix takes care of it by delaying the Firebug CSS loading. Fixes #4157.
      Replacing images with a new normalized set. Refs #3100.
      Created new dnd.css, images moved to resources, all DnD-related tests and demos are updated, Flickr Viewer is tweaked somewhat. Refs #3100.
      More tweaking to Flickr Viewer, removed gif files. Refs #3100.
      Removed English text from the default Avatar. Added an example on how to override the text using a custom code to Flickr Viewer. Refs #3100.
      Fixed typo in html.js. Tweaked the default DnD avatar for better visibility. Fixes #4210.
      Removed reference to dojo.lang.repr() --- it didn't harm, but we can save a few bytes by removing it. Thanks, guest! Fixes #4208.
      Now when dojo.dnd.Mover publishes a topic, it passes itself as a parameter, instead of a node. The node is still accessible from it as a "node" member. It gives more flexibility including the access to the latest mouse position, which is updated automatically on every move. Additionally subscribers can customize a mover on the fly.
      Splitting Mover and Moveable from move.js. Refs #4430.
      Moved topics from Mover to Moveable. Refs #4430.
      More moves from Mover to Moveable. Refs #4430.
      dnd: typo fix --- fixes #4799, thx jbondc for bring it to my attention.
      dnd: fixes #4140, thx blowery for the enhancement suggestion.
      dnd: fixes #4673.
      dnd: missed one place the first time around, added a test for start/stop events, fixes #4799.
      dnd: propagating changes from DnD to Moveable, refs #4673.
      dnd: propagating changes from DnD to Moveable, refs #4140.
      dnd: implemented optional drag handles for DnD items, added a test demonstrating the usage, fixes #4840, thx manishaNC for the suggestion.
      dnd: modified the behavior of the drop in case of moving items withing a source, fixes #4585, thx sarkasm for the suggestion.
      dnd: partial refactoring of dnd.move, which preserves the old API, but provides new facilities, refs #3695.
      dnd: constraints implemented with special Moveables instead of Movers, refs #3695.
      dnd: temporary removing files to change their case. Refs #4430.
      dnd: re-adding files back with the correct capitalization. Refs #4430.
      dnd: temporary removing two more files to change their case. Refs #4430.
      dnd: re-adding files back with the correct capitalization. Refs #4430.
      dnd: reflect changed capitalization of some DnD files. Fixes #4430.
      dnd: fixing missed renames. Fixes #4934.
      Implemented /dnd/drop/before topic. Fixes #5266.
      Reporting the overSource event only for the current source. Fixes #5268.
      Fixes #5332.
      More compact revision match. Refs #5332.
      Allow mouseless start of DnD. Refs #4140.
      Adding Date object support to dojo.clone(). Fixes #5284.
      dojo.fx: new versions of dojo.fx.combine() and dojo.fx.chain()
      dnd: minor cleanup (switching to dojo.deeclare for the sake of uniformity)
      Renaming onBeforeBegin => beforeBegin.
      Making sure that _fire() gets an array argument.
      Fixing one frequent case of dragging table rows without specifying
      fx: fixes #6018. !strict
      fx: Fixes #5539. !strict
      dnd: removing dojo.marginBox() calls from all places in the DnD package,
      dnd: removing an unnecessary comment. Refs #5687.
      dnd: adding moving "relative" objects. Updated the existing test
      dnd: removing unused debugging code (rather minor problem). Refs #6136.
      dnd: clean up stray naked dojo.query() statements. Refs #5687. !strict
      dnd: minor cleanup. Refs #5687.
      dnd: adding missing dojo.depricated(), refs #5687.
      dnd: removing the garbage, refs #5687.
      dnd: fixes #6273, thx haysmark!
      dnd: implementing the auto scroll for nodes. Fixes #1597.
      Added dojo.string.rep(), which creates a new string by repeating
      dojo.fx typo correction: fixes #6833. !strict
      dnd: clean up in Container. Refs #5687. !strict
      dnd: special treatmeant of the right mouse button on Safari/Mac.
      dnd: fixes #6923 (bleeding events), !strict
      dnd: fixes #6847, thx Revin Guillen! !strict
      dnd: adding the missing query restriction in the avatar, thx Bryan Forbes!
      DnD: adding a local event for a drag detection. Refs #7042. !strict
      DnD: adding local events to sources: onDrop, onDropExternal, onDropInternal.
      DnD: adding selfCopy flag. Thx, Revin Guillen! Fixes #7051. !strict
      DnD: making sure that the creator always called in the context of
      DnD: adding missing methods, adding synch() method to synchronize DOM nodes
      DnD: making the default creator to accept DOM nodes and process them
      DnD: added onDraggingOver and onDraggingOut local events, thx Revin Guillen!
      DnD: adding selfAccept for copyOnly sources, thx, Revin Guillen! Refs #7051. !strict
      dnd: adding the autoSync option, and the AutoSource class, which utilizes it by default.
      dnd: returning a context object from forInItems(). Thx, Max, for suggestion.
      dnd: removing some duplicated code. Thx, chucky! Fixes #7277. !strict
      Adding a delay parameter to DnD (just like in Moveable). Fixes #7436.
      Fixing a typo. Refs #5687. !strict
      Simplifying argument checks in dojo.place(), fixing a bug.
      Eliminating stray commas. Refs #7596. !strict
      dnd: preventing the selection on IE at the manager level, minor cleanup,
      dnd: fixing the DnD handles problem, !strict, fixes #7944.
      dojo.fx: fixing a bug in dojo.fx.combine(), thx ashendw!
      dojo.fx: simplifying dojo.fx.combine(), !strict, refs #6490.
      dojo.dnd: switching from isSafari to isWebkit, refs #7845, !strict.
      dojo.dnd: typo fix: switching from isWebkit to isWebKit, refs #7845, !strict.
      dojo._base.html: verified the problem, added the fix, modified tests,
      Implementing dojo.empty() with unit tests, !strict, refs #7976.
      Adding dojo.toDom() with unit tests, the code by James Burke (thank you!)
      Restructuring dojo.trim() and dojo.string.trim() to make a choice (native
      Reducing the code by several bytes and improving the load time a little, !strict, refs #8182.
      Adding more elements to the self closed tag dictionary, and to the wrap dictionary, !strict, refs #7890.
      Making the unique string scope-based, !strict, refs #7890.
      Adding dojo.toDom() to process dojo.attr(node, "innerHTML", frag) on IE correctly,
      Acting on James Burke's suggesting to simplify the placement of DocumentFragment, !strict, refs #7890.
      Implementing James Burke's suggestion to return a singular element directly without wrapping it in DocumentFragment, !strict, refs #7890.
      Taking out the HTML cleanup code, !strict, refs #7890.
      Removing fieldset-related records in the tag wrapping dictionary, !strict, refs #7890.
      Renaming dojo.toDom() to dojo._toDom(), updating all known referencing, !strict, refs #7890.
      Adding the HTML fragment support to dojo.place(), !strict, refs #7890.
      More comprehensive tests for dojo.place(), !strict, refs #7890.
      Implementing replace/only placing options with tests, !strict, refs #8380.
      Aliasing dojo.html._emptyNode() to dojo.empty(), !strict, refs #7976.
      Reusing dojo.create() in the base (back and dnd functionality), !strict, refs #7802.
      Adding dojo._toDom() functionality top dojo.create() with unit tests, !strict, refs #8446.
      Adding the handling of missing attrs with a unit test, !strict, refs #8446.
      Removing unnecessary notes from the inlined documentation, !strict, refs #8446.
      More inlined docs changes, !strict, refs #8446.
      refNode can be a string too --- making sure to count it in when attrs is missing, !strict, refs #8446.
      The return of dojo.place() is modified as agreed => now it returns the first argument resolved to a DOM node, !strict, refs #8380.
      Adding a missing unit test for a placement with a number rather than a string, !strict, refs #8380.
      Rolling back some changes to dojo.create(), updating docs and unit tests, !strict, refs #8446.
      Modifying the VML style rule to conform to Microsoft's guidelines. Refs #7617.
      Redirecting the "class" attribute access to use node.className, !strict, fixes #8499.
      Reshuffling mouse button processing, !strict, fixes #8603.
      fx: fixing a minor bug in dojo.fx.combine(), !strict, fixes #9302.
      html: cleaning dojo.attr() according to James' suggestions
      NodeList: correcting the unfortunate typo, adding a test for adopt(), !strict, fixes #9362.
      Removing unnecessary FIXME, !strict, refs #9362.
      Replacing "this" with "dojo", !strict, refs #9067.
      Making _everyOrSome a private function, !strict, refs #9067.
      Migrating the detection of Macintosh to hostenv_browser.js, !strict, refs #3740.
      Adding mouse-specific functions and a virtual key (a copy key), !strict, refs #3740.
      Switching DnD to new base facilities, !strict, refs #3740.
      Saving a few bytes by shortening isCopyKeyPressed to isCopyKey, !strict, refs #3740.
      Making public a list of DOM events used in NodeList, !strict, fixes #7974.
      We need to streamline all functions, which run as a part of the event processing, !strict, refs #3740.
      Revising dojo.attr() to adhere more to the original idea, !strict, refs #8991.
      Correcting typo in the SMD file, fixes #9436.
      dojo.delegate: removing a leak, !strict, fixes #9424.
      fx: fixing Toggler's docs, !strict, refs #9503.
      Eliminating stray commas, !strict, refs #9534.
      Changing the inline docs for dojo.colorFromArray + cleaning up trailing whitespaces, thx jjkress!, !strict, fixes #9415.
      Adding array support to addClass/removeClass, adding support to remove all classes,
      removing trailing spaces, !strict, refs #9560, refs #9577.
      dnd: minor tweak to improve flexibility of dragging with handles, thx Yoav Rubin, !strict, fixes #9593.
      dnd: minor tweak to dnd.css to change cursor over DnD handles, thx Yoav Rubin, !strict, fixes #9593.
      fixing the inline doc example for NodeList, thx Les!, !strict, fixes #9642.
      Fixing dojo._mixin() to support properties skipped on some browsers (IE) +
      Eliminating "this" from the loader modules, !strict, refs #9067.
      Small style change to eliminate "argument redefinition" warning, !strict, refs #9648.
      Minor speed improvements: eliminating "with", chaining "if"s, !strict, refs #9648.
      Simplification of dojo._mixin() + more unit tests, !strict, fixes #9409.
      dojo.clone() changes, !strict, fixes #9763.
      Simplifying one line in parser.js, !strict, fixes #9784.
      More code reduction, !strict, fixes #9784.
      Making sure that properties available on the empty object do not override custom properties on the target, !strict, refs #9409.
      Adobe Air uses a different key mapping on Mac, !strict, fixes #9840.
      declare: committing the new version with updated inline documentation, and getInherited(), !strict, refs #9862, refs #9795.
      dojo.declare: minor refactoring, support for dynamically redefined methods, !strict, refs #9862.
      dojo.declare: new test that uses 3 different forms of prototype method augmentation/replacement, !strict, refs #9862.
      dojo.declare: more accurate work with inherited "raw classes", !strict, refs #9862.
      declare: minor code reduction, !strict, refs #9862.
      dojo.declare: faster algorithm for C3 MRO, !strict, refs #9862.
      dojo.declare: a speed bust, !strict, refs #9862.
      dojo.declare: minor refactoring, !strict, refs #9862.
      Fixing the bug reported by Bill, switching to arrays for C3 MRO algorithm, minor speed tweaks, !strict, refs #9862.
      "constructor" was missing from the list of "invisible" properties (for IE), !strict, refs #9409.
      dojo.declare: fixing a bug reported by toonetown, thx!, !strict, refs #9862.
      dojo.declare: minor clean up, renaming, and speeding up, !strict, refs #9862.
      dojo.declare: different way to produce empty functions closure-free, !strict, refs #9862.
      dojo.declare: refactoring, size reduction, speed improvements, adding extend() like in the old dojo.declare(), !strict, refs #9862.
      dojo.declare: more refactoring to conserve space, speed up, and process edge cases, !strict, refs #9862.
      dojo.declare: more test cases with esoteric inheritance, !strict, refs #9862.
      dojo.declare: removing extra spaces, !strict, refs #9862.
      dojo.declare: speed up and size reduction, !strict, refs #9862.
      Removing trailing spaces, !strict, refs #9534.
      dojo.declare: replacing dojo.makeDeclare() with a special property, !strict, refs #9862.
      dojo.declare: performance improvements, !strict, refs #9862.
      declare: missed an initialization, !strict, fixes #10169, fixes #10170.
      Committing dojo.replace(), !strict, refs #9494.
      Comitting a patch to NodeList.concat(), thx John Ryding (IBM, CCLA)!, !strict, fixes #9337.
      declare: adding immediate parents to the meta information, !strict, refs #9862.
      Removing stray comma in the code example for dojo.replace, !strict, refs #9534.
      dojo.declare: performance improvements, minor API updates, !strict, refs #9862.
      dojo.declare: regression cleanup, !strict, fixes #10300, refs #9862.
      dojo.declare: fixing the kinks reported by Bill (thx!), !strict, refs #9862.
      dojo.declare: inlining and some code reduction, !strict, refs #9862.
      dojo.declare: removing unnecessary assignments, !strict, refs #9862.
      dojo.declare: cleanup, !strict, refs #9862.
      dojo.declare: making "constructor" a named variable, reformulating inherited() to gain back performance on Chrome, !strict, refs #9862.
      dojo.declare: eliminating trailing spaces, !strict, refs #9862.
      dojo.declare: fixing another gotcha by Bill (thx!), !strict, refs #9862.
      dojox.gfx: different initialization for IE8, !strict, fixes #8965.
      Renaming getffectiveAttrValue to getNodeProp as agreed on the meeting, !strict, refs #8991.
      Removing a stray space, !strict, refs #3961.
      Tiny style change, !strict, refs #3961.
      dojo.clonedojo.clone: changing the DOM node for IE, thx justinc!, !strict, fixes #9763.
      dnd: adding the missing function back, !strict, refs #10738.
      Fixing a special case for moving relative/absolute nodes without left/top offsets, thx ID!, !strict, fixes #10629.
      Using own properties instead of a shadow, !strict, fixes #10709, fixes #10788.
      Better troubleshooting in the case of missing base classes, stricter tests, !strict, refs #10829.
      Added the original event object to event handlers, !strict, fixes #10921.
      Cleanup, !strict, refs #9534.
      Added Tom's case, when base is an automatically created object, !strict, fixes #10829.
      Tweaking zoom on IE for opacity, thx CnEY, !strict, fixes #10409.
      ONly existing filter should be disabled, thx CnEY!, !strict, fixes #10408.
      Fixing the parents, thx rcgill!, !strict, fixes #10982.
      Added reuse of ids for DnD nodes, !strict, fixes #8863.
      gfx: bugfix - added missing shape elements to the list, thx Jared!, !strict, refs #10491.
      Performance improvement for dojo.addClass(), thx shaneosullivan!, !strict, fixes #11372.
      bugfix: VML initialization was wrapped in a closure, thx neonstalwart! !strict, fixes #11291.
      Added checks for constructor being called without "new", thx bradfordcarlsmith!, !strict, fixes #11237.

Fredrik Johansson (1):
      Allow setting a property to false from markup, Fixes #3832

Heng Liu (15):
      fixes #2877
      fixes #2878
      fixes #5313
      fixes #5421
      fixes #5452 (!strict, warning from exsiting content)
      refs #5452: handle undefined/null correctly when outputing them in console (!strict)
      fixes #5473 !strict
      refs #5506: change styles to style
      fixes #4613 and #6544
      refs #7178: added a test case for dojo._toArray with DomCollection
      fixes #7896: readonly needs to be treated different (just as tabindex) for IE<8
      fixes #10686: make sure null can be used to clear cache in dojo.cache
      fixes #10835: allow setting value to empty string in dojo.cache. added unit test.
      fixes #10975: expose event to subscriber of /dnd/drop
      fixes #11106: fixed an edge case problem for the new IE dom ready approximation

James Burke (278):
      Refs #2500. Tests for dojo.getObject()
      Refs #2500. Adding some tests for loader.js
      Refs #2500. Starting tests for hostenvs. Might not be working in rhino and spidermonkey
      Refs #2607. Fixing up baseUrl so it works right after a build.
      (merge from 0.4 branch) Fixes #2683. Loading local modules after page load should work now.
      Refs #2608: Change _baseUrl to baseUrl to allow developers to make url paths relative to it, also make Uri() constructor private, since we might want to change it to make it simpler. Waiting to see if there are use cases that require its current complexity.
      Refs #2608. Changing dojo.global() and doc() to properties for smaller code size and performance. No credible use case to have them as functions.
      Refs #2608. Get the unit tests in rhino to work, except for json tests. Rhino does not serialize object properties in the same order as they appeared in the source string. Probably something to do with for/in loops not guaranteed to happen in any particular order. We just get lucky in the browser. Need to possibly rethink the tests for it.
      Refs #2608. Fixes json tests so that they pass in Rhino.
      (merge from 0.4 branch) Fixes #2683. Did not fix debugAtAllCosts case with previous fix, and I was trying to fix it in the wrong way.
      Fixes #2747: _byId was defined after it was first used, and the dojo compressor stripped it out.
      Refs #2771. An edge case when loading local and remote modules via xd. Second part of fix comes over when loader_xd.js is ported.
      Refs #2911. Get jsonp callbacks working.
      Fixes #2989. preventCache is now supported for IO calls
      Refs #2911. Get jsonp callbacks working. Forgot a test file.
      Refs #2748. In process of porting loader_xd.js to 0.9. Not working yet (particularly for i18n)
      Refs #2990. Moving firebug out of base, but keeping it private. Confirmed with Alex.
      Refs #2990. Updating firebug console references to the right place.
      Refs #3237: changes dojo.io.script's jsonpParam to callbackParamName
      Fixes #3065, adds load, error, handle function callback support for io methods.
      Fixes #3240: adds guards around module code to avoid redefitions in layered build scenarios
      Refs #2748. Getting xd loader working. Not quite there yet.
      Refs #3252, matching the part in r9012 about ensuring that the _loadedUrls property of the dojo object actually acts as an array
      Refs #2748. Include dojo.i18n in dojo.xd.js. Otherwise, the logic was getting complicated to dynamically load it, and wait to process requireLocalization calls after it is loaded. With dojo.i18n included, dojo.xd.js gzipped is 27146 vs. 25477 without it (dojo.js is 23062 gzipped).
      Fixes #3340: Gets rid of an odd activexobject error sometimes showing up in firefox firebug.
      Fixes #3340: Second try. Previous attempt bypassed safari (thanks Bill for the catch). Gets rid of an odd activexobject error sometimes showing up in firefox firebug.
      Fixes #3388: Converted firebug lite to work as a div in the page instead of an iframe. Now usable in xdomain cases.
      Refs #3425. Updating version string
      Refs #2912. Starting port of IframeIO. Just doing a copy over first.
      Refs #2912. Porting dojo.io.iframe. Not there yet. Also made io arg handleAs understandable in xhr code.
      Refs #2912. Porting dojo.io.iframe. Not there yet. Got one very simple test to work. Still needs more tests.
      Refs #2912. Got an upload test working. Not part of doh unit tests since it requires user action to select a file to upload. It works in Firefox. Need to test other browsers and add an HTML response test.
      Fixes #2912. dojo.io.iframe ported. Tests succeed in IE 6, Opera 9.21, Safari 2.0.4 and FF 2.0
      Refs #744. Test for the window.eval issue in IE and Safari.
      Fixes #3500. Bad cancel() reference.
      Fixes #3509. Timeout path in general for io operations fixed, and specific issues in dojo.io.script were addressed. Thanks to dmachi for the timeout test file.
      Fixes #3139. Using the standard dojo.addOnLoad() to initiatilize the console div. Seems to fix the problem.
      Fixes #3534. args.url query params are not added to the list of params inside of dojo._ioSetArgs, but query params discovered as a result of dojo._ioSetArgs can be added to the querystring for qualifying http methods. The result is that the input args.url is preserved with its querystring params.
      Fixes #3577. Obsolete dojo.render reference.
      Fixes #3507: canceled deferreds get an error object with dojoType property of cancel.
      Fixes #3526. IO docs. Also updated port guide at http://dojotoolkit.org/book/dojo-porting-guide-0-4-x-0-9/io-transports-ajax
      Fixes #3588. Use new dojo._destroyElement call instead of removeChild
      Fixes #3702. Loading local i18n bundles was failing.
      Fixes #3978. Using dojo.doc instead of document inside of dojo.io.script.
      References #1140. In the timeout case, remove the script tag immediately instead of waiting for the next 50ms check. Some oddities with some browsers noted in the test page, but I will file bugs with the browser vendors once this test is in the nightlies.
      Refs #2319. Remove script elements in the jsonp style of usage that have already succeeded. This is not the fix though, and not sure if there will be a complete one without engaging browser vendors. Even in a test I did before without using dojo, just simple script attachments and removal of nodes when it was done, I saw a few bytes leaked on each call (I think the amount depends on the size of the response in the script).
      Fixes #3733: Exposes semi-public method to cancel all pending IO requests.
      Fixes #3724. Now IE 6 should pass xhr unit tests. Note that only testing from a server is officially supported, given that the tests try to do PUTs and DELETEs, which do not make sense for local files.
      Refs #3733. Need to put the function definition above the call, otherwise bad things happen with shrinksafe.
      Fixes #2765: calling dojo.addOnLoad inside a dojo.addOnLoad after doing an xdomain dojo.require() should work now.
      Fixes #3743. Need to wait until after browser_hostenv does script tag parsing for djConfig. Doing a copy/paste job in all hostenvs to avoid bloating base by having duplicate check in loader.js.
      Refs #2713. Should have been part of previous commit.
      Fixes #2713 #4085 #3801. Flattened bundles for i18n layers should work in xd and non-xd load cases now. Also brought over some xd fixes that were made after the 0.9 code was copied from the old trunk.
      Fixes #3984: adds a djConfig option to turn off the json-comment warning when using handleAs: 'json' with the xhr calls. Changed the handleAs names for dojo.io.iframe to be consistent with xhr handleAs names. Using dojo.fromJson in the dojo.io.iframe 'json' handleAs logic branch. Added some more unit tests for dojo.io.iframe.
      Fixes #3973. Support BASE tags. Does not work with Opera 9.22, but it could be a general issue with Opera, base tags and XHR loading. Still would like to fix the core issue of why we need a dojo._Url call in here, but can't seem to crack it.
      Fixes #3845. Needless double transform of form info, but also pointed to an issue with dojo.queryToObject, it was not calling decodeURIComponent for the query names. Fixed that too.
      #Fixes #4195. addOnUnload not being called. Resulted from r10151. Adding a simple test and cleaning up docs (thanks to liucougar for pointing that out).
      Refs #4183. Porting XHR IFrame Proxy. Cleared core change with Alex.
      Refs #4183. Porting XHR IFrame Proxy. Need to pass args object, not URL.
      Refs #4224. Removing document.domain references. Tested on FF 2.0.06, Safari 3.0.3, Opera 9.23, IE 6 and IE 7.
      Fixes #4288. Bad variable reference from a refactor. Fixes loading if the page uses an xdomain baseUrl and relative paths to that for the modulePaths.
      Refs #4331. Adding conditional comments to strip out functions duplicated in dojo.xd.js.
      Refs #4398. Add LICENSE file.
      Refs #4399. Rearrange code to separate loader concerns from other things. Ends up with a slightly smaller dojo.js too. Still a bit more to do on it, but this tested well.
      Fixes #4259 and #4330. Fixes for form values with multiple names and for selects with multiple selected options
      Fixes #3583. Introduces debugAtAllCosts support via dynamically loading loader_xd.js and a debug support file. Requires use of dojo.addOnLoad() to work.
      Fixes #4327. Single quotes in the value where causing trouble. Patch from schallm, CLA on file. Thank you.
      Fixes #4495. The extra set () should avoid the jslint error.
      Refs #739. Have xd loader use same comment removal code as the build system, so behavior is consistent. Also makes xd loader smaller.
      Fixes #3944. Prevent bad code that adds things to Object.prototype to cause errors in xdomain loading. Not bullet-proof, if an Object.prototype property is added that is a boolean true, xd loading will fail, but hey, don't do that to Object.prototype.
      Refs #4499, removing unused exactOnly argument to dojo._loadModule()
      Fixes #4938. Removing .* to __package__.js support.
      Refs #3961. requireInto is not ready for prime time yet. In particular, it will not work with xd loading. Removing for now, to help reduce base.
      Refs #4820. Format the kwArgs docs for IO methods so they show up nice in the parsed docs.
      Refs #4820. Try to avoid the user from hitting a bundle loading error case.
      Fixes #4978. Remove duplicate mousemove, and reorder names alphabetically.
      Fixes #5131. Make sure if trying to access parent frame, catch any exceptions and eat them. This is important if the iframe is on a different domain than the parent frame. (Checkin for trunk)
      Fixes #5367: dojo.isSafari was not reporting 2 if using Safari 2. (merge to trunk)
      Fixes #5408. Useless percent, thanks to peller for the catch.
      Fixes #5091: debugAtAllCosts including files twice in the head.
      Refs #4573: putting in scope changes. Say hello to _scopeName for string IDs. \!strict
      Refs #4573. Moving scope tests to their own directory
      Refs #4573. Moving scope tests to own directory, adding a djConfig burn in test.
      Refs #4573: Convert to dojo.config instead of djConfig internally. \!strict
      Refs #4573: get debugAtAllCosts to work with scope changes.
      Fixes #4573: Removed the path remapping in loadModule, since it messed up the ability to make your own namespace for dojo, but add your own code to it outside the dojo directory. Also finished book documentation.
      Fixes #5732. Allow other non-word character terminators like \; for dojo url.
      Fixes #5420. Allow other HTTP methods. To do a HEAD call: dojo.xhr('HEAD', args)
      Refs #5420. Accidentally committed this file. Reverting it.
      Refs #5420: goofed on the closure variable for dojo.
      Fixes #5767 (debugAtAllCosts working with dojo svn source) and #2904 (attaching cacheBust parameter inside of dojo._getText -- this allows dijit templates to be updated on reload if you set up djConfig.cacheBust correctly). \!strict
      Fixes #5838: allow console to work with xhtml documents. \!strict
      Fixes #1704: using DOMContentLoaded now for Firefox 3, since the hang bug is now fixed (as tested in FF 3beta3).
      Refs #5887. Introduces djConfig.afterOnLoad: if set to true, then dojo._loadInit will be fired after a 1 second timeout. Using a shorter timeout (100ms) had problems where the dojo._loadInit was fired before all modules were downloaded. That seemed odd given the synchronous XHR calls. Also because of a bug in the dependency mapping stuff, had to rearrange how _base.js is organized. Testing on supported browsers looks good, although Opera in the xdomain case seemed to have a problem. [...]
      Refs #5887. Gets opera to work after page load in xd situation. Opera was downloading modules as part of the dojo.config.require section, which happened before dojo.i18n was downloaded in the xd case. Switch to using nifty conditional include build comments so we can load dojo.i18n for dojo.xd.js before the dojo.config.require section runs in _base/browser.js \!strict
      Fixes #5576: allow args.form for dojo.io.iframe.send() calls to be a string that uses dojo.byId to find the node. Thanks rcgill for pointing out the API inconsistency.
      Refs #5887: previous fix is not so good: means that svn source usage of dojo.js will load dojo.i18n by default. Backing out that fix.
      Fixes #5801. Send X-Requested-With: XMLHttpRequest header with dojo.xhr calls, to match behavior of other libraries.
      Refs #5878: initial jaxer support. The oncallback() usage of Jaxer does not work yet, but the normal page request version does. \!strict for dojo.js
      Refs #6064. Allow dojo.io.script to use a child iframe. This gets the basics in there, and leaves the higher order issues of setting up the iframe to the user. Still need to document it. Ideally also show an example, but that might take longer.
      Fixes #6119. Make sure only do Jaxer server stuff if we are on the server. Thanks pottedmeat. \!strict for redeclarations
      Refs #5962. Apparent typo in docs for dojo.delegate \!strict
      Fixes #6274. Removing this test: it was for #744, but that is marked as wontfix.744
      Refs #6276. Allow multiversion support. \!strict
      Fixes #6270: back-bookmark.html was not ported from the 0.4.3 codebase. Removing it for now, not sure if it is still relevant anyway.
      Refs #6276. Make sure the attribute IDs are multiversion aware. \!strict
      Fixes #6203. Patch and unit tests from doughays. IPv6 URL support.
      Fixes #6525: all xhr calls go through dojo.xhr now. Thanks to Kris Zyp for the patch. I applied it with a slight mod to the postData and putData detection. Nice patch that reduces the code footprint, thanks Kris! Also fixed an incorrect test.
      Fixes #5887: move afterOnLoad kickoff after the dojo.config.require work in dojo._base. (merge to trunk)
      Fixes  #6358: xdomain loader did not handle module paths starting with /. Matching loader.js behavior and removing some cruft. \!strict for redeclarations (merge to trunk)
      Fixes #3242. Applies patch from Sam Foster. Removing weird IE WebControl logic that caused some race conditions with unload firing in IE. Updated test file to reflect test results.
      Fixes #6688. Allow spaces between provide and parenthesis. Also removed some redeclaration complaints coming from rhino strict checker
      Fixes #6380. Applying patch from Kris Zyp to deprecate json-comment-filtered. Updated 1.2 release notes too.
      Fixes #6380. Forgot to account for json-comment-optional. Thanks to gregwilkins for tracking down the issue.
      Fixes #6539. Non-string uris (like dojo._Uri objects) cause problems with cacheBust. Thanks to schallm for pinpointing the problem.
      Fixes #6727. Support a djConfig.addOnLoad option. Also updated API docs for djConfig.
      Refs #4462: Get xdomain loading to work for dojox.gfx. Requires a new loader function, dojo.loadInit() for this to work. Not my favorite, but it is the only robust option at this point.
      Fixes #6863. Only use try/catch in the djConfig.isDebug = false case. Otherwise, let the original exception escape to allow easier debugging. \!strict for _base/fx.js (error on line unrelated to changes for this ticket)
      Fixes #7052. the _listeners array was effectively a public, global value. Making it unique per dojo._scopeName. Reviewed change with sjmiles. Tested on IE 6. \!strict
      Refs #6064: get script removing working with frameDoc, and also fixes a multiversion bug with a 'dojo' string reference.
      Refs #6863: small try/catch error reporting cleanup.
      Refs #1368: make sure console calls are still useful outside of built files that have the calls stripped.
      Refs #1368: make sure xdomain warnings are really warnings. \!strict
      Fixes #7114: multiversion support for dojox.gfx/dojo.requireIf in xdomain loading.
      Refs #6793: Expanded error callback documentation to reflect recent changes to try/catch behavior.
      Fixes #7119: customBase now injects dojo._base dependencies. Includes standardCustomBase build profile that can generate the smallest possible dojo.js needed for dojo to work. dojo.js with that profile ends up at 13,690 bytes shrinksafed, 5,558 bytes gzipped. dojo.xd.js is 25,387 bytes shrinksafed, 9,101 bytes gzipped. Required a modification to dojo.i18n to trick the customBase regexps to get that dojo.xd.js size. \!strict.
      Fixes #5180: move gears detection out of Dojo Base. Now in a dojo.gears module. Tested change in FF3 with Gears, using dojox.off and dojox.storage demos. \!strict
      Fixes #6909: xhr calls were not being aborted onunload in IE7 since typeof xhr.abort is 'object' in that case, and not 'function'.
      Fixes #6821: Preserve exception info in case where xhr.send fails. The original exception is accessible off the response.cancelResult property inside the handle(response, ioArgs) callback.
      Fixes #6786. Introduces dojo.addOnWindowOnload and ties IE leak cleanup to that method. Allows file download links and javascript links to work in a page without destroying widgets. \!strict
      Fixes #7233. Fixes dojo.io.iframe docs, and also adjusted dojo.io.script's docs to match.
      Fixes #7250: reuse previous response if no new response from callback. Thanks pottedmeat!
      Fixes #7387, order of form elements off in dojo.objectToQuery, and Refs #6821. My fix for 6821 dropped the erro.dojoType = cancel which caused xhr doh unit test to fail.
      Fixes #7435, optimize console calls in non-debug case. Thanks to jcerruti (CCLA on file) for the patch and testing coverage.
      Refs #7496: Make sure dojo.io.script documents the io args in the same way dojo.io.iframe and dojo.__xhrArgs does too, so that dojo.__IoArgs shows up as parent.
      Refs #7523. Use conditional build options to leave out rhino-specific fix for browser builds, to limit impact on dojo base size. \!strict
      Fixes #7667. Thanks for dfabulich for tracing down the problem and providing a patch to indicate what still needed to be checked. I did not use the patch as-is though: from what I can tell the extra setTimeout call to _xdDebugFileLoaded is not needed if we clear the currentResourceName before doing the check for inflight modules.
      Fixes #7714. The change to switching to setting Enabled on a filter (to preserve other filters on the node) left a previously set value of opacity as the opacity value even though it was not active. Make sure we update the opacity value to 1, but still make sure the filter is set to Enabled = false for the opacity == 1 case. \!strict
      Refs #6876: a test page for keypress events.
      Fixes #6876. Thanks to Bill for improving the patch. Arrow keys now detectable via onkeypress in Safari 3.1 without using IE code paths. Also, dojo.disconnect for onmouseenter/leave for non-IE browsers should work now. \!strict for existing code patterns.
      Refs #7330. My OCD compels me to alphabetize these names.
      Fixes #5574 and Fixes #7696. dojo.isChrome and dojo.isWebKit available. dojo.isWebKit now gives version numbers like 525.3. KHTML detection changed to just support being defined for KHTML browsers only, since other KHTML-related browsers are likely more sensitive to dojo.isWebKit versions. Changed dojo.isSafari check so that now Safari 3.1.2 now correctly reports as dojo.isSafari 3.1. But with that change, Chrome browser does not have dojo.isSafari defined, which should be good, sin [...]
      Refs #7696. Cleans up docs on isKhtml.
      Refs #5692. Additional docs for dojo.style. \!strict
      Fixes #8103, favor using native browser console instead of firebug lite. Helpful in IE8. Also includes docs for a couple more djConfig parameters. \!strict
      Fixes #8072: firebug lite interfering with dojo.query tests. noFirebugLite: true was being used, but according to _firebug/firebug.js it is deprecated now in favor of just isDebug: false. Also changes in 8103 to use native consoles also reduces the need for firebug lite for this test page.
      Fixes #8048, Firebug lite broken after shrinksafe runs over the file. Happens because of a bug in shrinksafe where it does not properly replace vars that are placed after they are first used. \!strict
      Refs #8103: rolling back changes to use native console as much as possible. Turns out at least IE and Safari do not have console.debug at the very least and do not support console.dir and some other functions. I should have tested more pages than my simple test case used for the previous change. \!strict
      Fixes #7981: move the throw-instead-of-catch behavior for deferreds and related to trigger off of debugAtAllCosts. \!strict
      Fixes #8023, allow for more descriptive cancel messages in Deferred.
      Fixes #8270: Chrome: doojo.coords incorrect value on width. \!strict
      Refs #7845, scrubbing dojo.isSafari use in Core, using isWebKit where appropriate. Have not done dojo.dnd internals though, and dojo._firebug had some isSafari references too, but not sure those need to change. \!strict
      Fixes #5732, allow dojo.js urls to end in a semicolon. Using the same match as used in hostenv_browser.js. \!strict
      Fixes #7875, make sure djConfig.require path uses multiversion scope name. This approach also works better in non-eval friendly envs like AIR.
      Fixes #8310, dojo.attr should work with defaultValue. \!strict
      Refs #8310: rolling back r16125, there is a larger question about supporting fetching properties of a DOM element vs things via getAttribute() and things that are considered true dom attributes. Adding all of those dom properties to _attrProps is not scalable. \!strict
      Fixes #8394, odd query issue in IE8, fixes tests._base.html_element tests in IE 8. \!strict
      Refs #8394, another query test for the weird case, but seems to pass in the query doh tests even before the 8394 fix.
      Fixes #7690, dojo.byId fix, thanks to Mike Wilson for the pointers from DWR. Unit tests added too. \!strict for existing code
      Fixes #8495 using patch from doughays. Thanks for catching the bad previous fix.
      Refs #8467: minimize xhr callbacks happening in the middle of an execution thread when starting a new XHR call. Still needs more work for sync requests, so leaving ticket open for a future milestone to address sync case.
      Fixes #8047: IE8rc1 is much better now, and we do not need an offset calculation for nodes. There is still an issue with rtl direction and the offsetLeft, but will open a separate bug on that, but I believe it is Microsoft's to fix, it is a strange negative number. \!strict
      Refs #8526, some small doc changes.
      Fixes #8506: adjust dojo.isIE if x-ua-compatible EmulateIE7 is used.
      Refs #8543: doc update for note about pixel values. \!strict
      Refs #8547, allow sizzle as a build option replacement for normal dojo query engine. Still some rough edges/behavior differences, and want to be sure this is the best way to integrate. \!strict
      Refs #8547, get the sizzle wrapper to do more of dojo.query-like work when query is not a string, then for both acme and sizzle, make sure the dojo detection works with multiversion support. \!strict for sizzle code
      Fixes #8598: change var listCtor to dojo._queryListCtor. Also had to move getNodeIndex above first use to avoid shrinksafe bug with mangled names. \!strict
      Getting rid of console.debug calls even in comments/doc comments for code in Core, in an effort to move away from console.debug, since it is not supported in other native consoles like webkit and IE8. Refs #8583 \!strict
      Refs #8612 and #8613. Reverts changesets [16641] [16642] [16643]. dojo.create not accepting html strings has been explicitly discussed as not being included in Dojo 1.3 multiple times now. We can talk about it more for 1.4 after we have had time to let the current create bake a bit in the wild. It is easier to add things later than take away later too. Similarly with NodeList, while I want an end() in general, i want to be sure we think through the use cases for it, and adding more  [...]
      Fixes #8621, dojo._toDom auto-selecting an option tag. \!strict
      Fixes #7633, upload.cgi had a bad percent that was not escaped. Also turned off file saving by default to avoid common permission/directory missing errors.
      Fixes #8282, applying Eugene's patch with expanded unit tests for NodeList.addContent
      Fixes #8744. Thanks klipstein for tracking down the problem with multiversion suppport.
      Fixes #8797, useless code.
      Refs #8730, mention dojo.place as a viable alternative to dojo.html.set.
      Fixes #8839, edge case with ID selectors that have a root. Updated unit tests. \!strict
      Fixes #8895. Remove accidental isGecko, and hack patch for firefox variants. Not a complete detection but better than what we do now, and least risk. Need a better solution for 1.4
      Refs #8895, slight code reduction.
      Fixes #8906: dojo.byId regression. Why IE do you make it so hard to get an element by ID? \!strict
      Fixes #8940: IE8 now includes the parent's border in offsetLeft/Top \!strict
      Refs #8940, missed the quirks box test failure. \!strict
      Fixes #8076: IE8 no longer needs the bidiScrollLeft fix, neato. \!strict
      Fixes #8937, get isFunction to work for strange Opera bug. We shuld still follow up with Opera though and report the problem but at least we can work with today's opera. Once they fix, we can remove the separate var assignment. Applying patch from peller.
      Refs #8938. Fixes XML unit tests for query in Opera. We were not setting caseSensitive in that case. The doctype and contenType do not give any value, so reverting to old school toString hackery. Should be OK given that it is limited to the opera case, but subject to review from Alex. \!strict
      Refs #8906: a cleaner fix for cloned nodes that has their IDs changed. Added test case for testing the 'form with child element that shows up as an id property on the form.\!strict
      Refs #8598: restore capability in query.js to dynamically change the node list constructor for each query call. \!strict
      Fixes #9001: clarify docs on the purpose of dojo.exists.
      Fixes #9172, content with array property was not being cleaned up properly and IE the code threw an error.
      Fixes #8732, adds script onload detection. Most of the core trick taken from plugd, from phiggins. Also allows jsonp instead of awkward callbackParamName for setting jsonp callback name. \!strict for some bizarre no side effect warning.
      Fixes #9219, #8649, #7833 and #8724: allow publication of IO pipeline topics, add dojo.fieldToObject (thanks to foobarfighter, CLA on file), allow generic args.rawBody for setting HTTP request bodies, and if hasBody is not passed to dojo.xhr, make an intelligent guess as to whether the query should be in the http body or as a query string. \!strict for dojo.io.script code has not side effects warning, but it is ok.
      Doc update, Refs #9219
      Fixes #9133: improve json-comment-filtered. Thanks to doughays for the basic patch: looks like our unit tests allow for other code, like while(true); block before the start of the comment, so we need to look for explicit starts of an object or array.
      Refs #9250: allow dojo.parser to expose the attribute used for querying, so that they can be reset for widgetsInTemplate parsing in multiversion scenarios.
      Refs #8612: re-introduces a _stash and end() method for dojo.NodeList. Also, methods that return new NodeLists also call _stash() so end() can be used properly with them. dojo._queryListCtor renamed to dojo._NodeListCtor since it is a more accurate description and also set by dojo.NodeList. This allows other dojo.NodeList delegates/variants to be used by our query and NodeList operations. Introduces traversing methods for dojo.NodeList via dojo.NodeList-traverse: children, closest,  [...]
      Refs #8612: extra checks for null parents (suggested by uhop). parent() really just needs to get immediate parent, and from what I can tell, element nodes are not embedded in anything hazardous except document nodes, but those are weeded out in _getUniqueAsNodeList. Also added a doc clarification to clarify name of function.
      Refs #8612: better docs for _stash and end, code reduction in NodeList-traverse methods. For NodeList.end, throw descriptive error if no parent NodeList. This is experimental though/needs more input. Pete for one is not sold on the idea, but I feel we need to fail as fast as we can instead of silently introducing bugs into a new user's code.
      Refs #9341: dojo.cache implementation.
      Refs #8612. More NodeList sugar in the form of manipulation APIs for placing nodes in different positions. This one is a bit interesting because it will define an html() method. So does dojo.NodeList-html. The one in -manipulate will defer to the one in dojo.NodeList-html if it is loaded. html() in -manipulate is just an alias for -manipulate's innerHTML method, so that one can always be used if you want to be totally safe. I did not want to use -html's version of html(): I feel it  [...]
      Fixes #9353: Fix from Mark Wubben (CLA on file) for a better node test that works in other environments like greasemonkey. \!strict for existing code
      Refs #8612, now addContent can accept HTML string templates and parse for widgets, if dojo.string and/or dojo.parser have been separately dojo.required. This is a bit experimental. If the feedback from folks is good, it will stay in and get more docs on dojocampus. These changes along with dojo.cache allow for some neat NodeList-based HTML construction from html templates. It is tempting to put in support for dojoAttachEvent and dojoAttachPoint, but I think with the nodelist operati [...]
      Refs #9359: clarify what kinds of queries can be used with the traverse methods, basically simple ones, no decent operations. May modify it later if found to be too restricting. Opting for speed in this case until viable use cases otherwise are identified.
      Refs #8612: updating docs to mention other position options that became available as of Dojo 1.3
      Fixes #8976: patch from doughays, thanks for the patch and test case.
      Refs #8606: expand dojo.require docs to talk about async loading.
      Refs #8612. Inline doc examples for dojo.NodeList-manipulate
      Refs #8612. Inline doc examples for dojo.NodeList-traverse
      Refs #8612: allow for widgets that change their DOM node.
      Fixes #9402, object element returns true for dojo.isFunction() in Firefox.
      Refs #8612, fixes instantiate() error and makes sure test passes on all browsers.
      Fixes #9420, #8649, treat single value radio or checkboxes as just a string. Thanks to Maine (CLA) for testcases and the fix.
      Fixes #6912, #9418. Removed unused libraryScriptUri and _rhinoCurrentScriptViaJava, introduced dojo._getText for rhino so that dojo.cache now works in the unit tests under rhino. Firefox extension hostenv already has a _getText. Spidermonkey does not, but I expect more is missing from that hostenv and so not messing with that one for now. Also corrected some strict errors.
      Refs #8612. Remove instantiated from NodeList. Getting dijits works via code snippet in the ticket. it does not work for generically declared things, but it is not clear instantiated is useful in NodeList type use cases.
      Refs #8612. Be sure the NodeList-manipulate methods that use _placeMultiple can handle NodeLists and nodes as arguments.
      Fixes #9452, dojo.isDescendant failing in GreaseMonkey scripts. \!strict
      Refs #8612. Doc typo fixes.
      Refs #8612: NodeList sugar: make it easier for dojo.NodeList delegates to use their construcors for any NodeLists created while part of chained methods.
      Fixes #9584: use charset for xd script tags, allow for changing it, but not sure if that private option is strictly needed. Seems like the developer would have to go through some hoops to allow loading a non-utf8 module/building those modules.
      Fixes #9659, a debugAtAllCosts race condition.
      Fixes #9649, prependTo/appendTo returning wrong values. Also added returns: comments for all -manipulate and -traverse methods. Fixed an issue with latest firebug throwing off the -manipulate and -traverse tests.
      Fixes #9662, a bug with wrapAll.
      Fixes #9727, where a dangling interval might be kept running in the xd loader case.
      Refs #8606, doc clarification for dojo.fromJson
      Refs #9341, allow dojo.cache to accept objects with .toString() methods that return a valid path to the text file that should be cached.
      Refs #8606, docs to mention using every or some instead of forEach if want to break out of the loop.
      Refs #8606, docs to mention using every or some instead of forEach if want to break out of the loop.
      Fixes #9839, missing test file.
      Fixes #9816, bad code in i18n handling in xd loader. Only triggered by bizarre code loading sequence.
      Fixes #9895, informs developer if dojo.back.init() is not called at the right time, avoiding a document.write that would wipe out the page and be hard to debug. \!strict
      Fixes #9925, dojo.ready as an alias for dojo.addOnLoad. Docs, release notes updated too.
      Refs #9942, adds a note to dojo.forEach about operating over sparse arrays, and updates devmo links for the different array methods.
      Fixes #9672, nextAll, prevAll and andSelf added to dojo.NodeList-traverse, with docs on campus.
      Fixes #9434, remove bloat code from xhr.js, adjusted unit test to do something more realistic to test xml content handling.
      Fixes #9975, taking out FF2 branch since FF2 is no longer supported. Technically, does not fix the issue in the ticket, but resolves the ticket by saying it is now an unsupported browser case. \!strict
      Refs #8612, forgot to remove a line when I removed the instantiated stuff.
      Refs #8612, update end() docs to match code.
      Fixes #9495: skip deferreds that have been canceled. Used a different fix than the ticket patch, dfd.canceled is the right spelling, and it seemed best to do the work in the fireNext function.
      Refs #9434 and Refs #9220, had to roll back removal of IE xhr handling code in order to get the dojo.io.iframe tests to work. To bad we cannot remove the xhr code, in retrospect, probably better not to support xml handling in dojo.io.iframe -- it sucks, and there is no universal solution (Jared found that he needed different hacks for very large XML).
      Fixes #9050, use dojo.doc where applicable.
      Refs #9717, documenting that nls is an unsupported bundle name. Leaving ticket open until docs.dojocampus.org is updated (once the server recognizes login).
      Refs #9183, changing last isFF references to isMoz for Core. there is one isFF reference still in use, but it is to specifically detect for a bug specific to FF 3. \!strict
      Fixes #9486, enable not sending X-Requested-With by setting that header value to an empty string on a per-request basis.
      Fixes #9672, documenting that prevAll returns nodes in reverse DOM order. This matches the current jquery 1.3.2 behavior.
      Applying patch from Rob Retchless, CLA. Fixes #9969
      Fixes #9311, nested addCallback calls end up with the nested callback not getting the right, up to date response.
      Fixes #10081, make the DOMContentLoaded simpler and do not rely on a script defer in IE.
      Refs #10081, put back in the script defer for IE. David Mark pointed out that document.readyState changes at window.onload time, so it basically means previous change only fired addOnLoad callbacks in IE on window onload. While the edge case breakage with the script defer is unfortunate, losing DOMContentLoaded for the majority case in IE is more unfortunate. So putting it back. But removed the interval timer since it was basically useless.
      Fixes #10021, use more generic tests for documentMode, but seems like a very weird test scenario.
      Refs #10029: fixes the children() method and all NodeList-traverse methods to use a slower dojo.query-based filter vs dojo._filterQueryResult. It allows more types of CSS selectors. Updated dojocampus docs for -travers. Leaving the ticket open to figure out what to do for dojo.NodeList's .filter method, but will be pushing that out to another milestone.
      Fixes #10159, djConfig.afterOnLoad only works with built dojo.js files, so noting it in the documentation. \!strict
      Fixes #10174, remove this references for better performance. Fixed a multiversion test. \!strict
      Fixes #10139. Thanks to Les and elazutkin for doing the investigation. Using elazutkin final suggestion. Checked unit tests in browsers and in Rhino and they pass with the change.
      Fixes #10207, help IE help itself.
      Fixes #10208, leak patch from Bill for Firebug. \!strict
      Fixes #8732. It would help if I actually check in test files.
      Fixes #10292, djConfig value for skipping the IE DOMContentLoaded hack. \!strict
      Fixes #10287, use the shorter ioArgs where possible, expand the comment about the _xhrObj call.
      Fixes #10259, FF 3.6preb2 reports itself as Namoroka and not Firefox, but do not want to build in extra logic to isFF since we just got rid of needing it except for backwards versions, so fixing firebug.js by making sure it does not try to do the console work if firebug's console is defined (known via console.firebug) \!strict
      Fixes #5179, thanks to sfoster for doing the work figuring it out. Fixes :checked for option elements \!strict
      Fixes #8358, some infix operators do not require spaces around them. Tested with Alex's queries/mss tool and normal slickspeed and do not see appreciable differences with this change. \!strict
      Fixes #10142. Thanks to arv for providing a test case that identifies the problem and identifying the backwards loop. \!strict
      Refs #8358, patch to my patch from doughayes that is more compact and prettier. Tests still pass, and no noticeable drag on perf over previous patch. \!strict
      Refs #8606, add caveat to dojo.back about semantic frag IDs. \!strict
      Refs #8606, clean up some of the docs for dojo.eval \!strict
      Refs #8606, clean up some of the docs for dojo.eval \!strict
      Fixes #10809
      Fixes #10918, move away from old IE domcontentloaded approximation to something that should work better. Made detection of page loaded after dojo is added to the page better too.
      Refs #10918, make the IE interval cancel more robust.
      Fixes #10980, build and xd loader choke on the dojo.require() reference, even though it is in a string. the regexps used for the build/xd loader are string-aware. Technically not dojo.declare fault, but this is the easiest fix. \!strict
      Fixes #10824, from patch, CLA on file
      Fixes #10628, helps out IE memory management, thanks to shaneosullivan for confirming the simplified fix.
      Fixes #11074 doc typos.
      Refs #7757 and Fixes #11090, do not throw an error for just running dojo.js from source in firefox with firebug gbreaking on all errors. Start the migration to hostenvs explicitly indicated by the user as it will be increasingly hard to detect the correct hostenv anyway. \!strict
      Fixes #11230, allow dojo loader to work in chrome extension.
      Refs #10728, update docs to mention that dojo.contentBox cannot set left and top \!strict
      Fixes #10814, some globals were linking, applied patch from markwubben. \!strict
      Fixes #11510: fix issue with regexp looking like a comment for debugAtAllCosts removal.
      Refs #8546, allow using protocol-relative URLs with xd loader (backport to 1.5 branch)

Jared Jurkiewicz (98):
      Minor update that fixes #2801  Doc updates for clarification and a new
      Simple update to UT.  refs #2845
      Fixes to minor issues such as a var name consistency issue and a spacing nit.  fixes #2891
      Checking in application of Michael Smith's #2542 sorter patch + UT +
      Applying two patches from Michael Smith for minor documentation updates (dojo.data.util.simpleFetch_20070429.patch and dojo.data.api.Read_20070501.patch).  refs #2542
      Adding in queryIgnoreCase option into JsonItemStore (from dojo.data meeting on 2007.05.01 )  refs #2909
      Minor update to JsonItemStore tests to remove functions from the global space.  refs #2950
      Simple conformance tests for JsonItemStore.  refs #3109
      Committing in minor updates.  refs #3185
      Committing in fix for warning about comment filtered Json in JsonItemStore.  fixes #3098
      Fix to handle null cases, as well as a performance fix + UT.  fixes #3153
      Putting in performance patch.  refs #3261
      Updated API files for dojo.data to remove use of dojo.unimplemented.  refs #3220
      Commiting in minor change to remove the attribute-item concept from the API.  fixes #3291
      Checking in minor API addition.  refs #3207
      Adding in getLabel API. refs #3299
      Minor doc updates for identity.  refs #3358
      Trivial cleanup.  fixes #3372
      Removal of now unused test data.  fixes #3362
      Updating UT for JsonItemStore.  fixes #3361
      Minor tweaks to the Notification docs.  refs #3358
      More minor doc updates.  refs #3358
      API update for dojo.data.api.Identity.  refs #3554
      Checking in fix to the date serializer to use dojo.date.stamp to use ISO formats for storage and default serialize/deserialize.  fixes #3679
      Update to testcases to get around a problem with Rhino running the UT.  Further investigation should still be done....  fixes #3729
      Fixing minor issue with RPC tests and Rhino.  fixes #3730
      Updates to custom data type handling.  fixes #3792
      Trivial fix in ItemFireWriteStore fixes #3824
      Patch to 0.9 for form handling issue.  fixes #2844
      Fixing browser host env to get local file loads working on IE 7 (and providing option to force ActiveX xhr all the time if so desirred.)  refs #1990
      Adding a small bit to the Write API.  fixes #3862
      Minor doc update that was missed. refs #3862
      Deliberately trying to commit a file that won't pass UTF-8 checks.  If this gets in, the check doesn't work.  refs #3938
      Checking if fix to  non-tf8 char.  Byte emcodin check didn't catch the bad bytes.   Fooey, validator test failed.  Restoring to good file.  refs #3938
      Removing accidental global.  refs #4066
      Updates to handle multiple fetches odduring at data load time.  refs #4067
      Minor fix to save.  fixes #4394
      Minor fix.  fixes #4481
      Another minor ref fix.  refs #4394
      Minor fix to sorter.  fixes #4491
      Updating error message.  fixes #4629
      Fix for identity problem.  Fixes #4691
      Minor fix to squelch warnings,  refs  #4829
      Committing in minor enhancement for RPC.  fixes #4888
      After more looking, I don't think you can do jsonp comment filtered.  So reverting #4888.  refs #4888
      Couple UT.  fixes #4931
      Minor patch.  fixes #5307
      Minor defect fix + UT.  fixes #5357
      Fix for a fat-fingered typo caused by #5399, 'iobject' instead of 'object'.  refs #5399
      Updated to ItemFileReadStore and ItemFileWriteStore to enable reference integrity.  fixes #4552 \!strict
      Fix to ItemFileWriteStore for reversion bug.  refs #5743 \!strict
      Fix comment typo.  refs #5743 \!strict
      Minor fix.  refs #6016
      Addition of preventCache into ItemFileReadStore.  fixes #6072 \!strict
      API clarification refs #6564
      Added in support for reload when close() is called.  fixes #6073 \!strict
      Adding in fatal error for a situation that cannot be worked around (as far as I understand what could be done with JS)  fixes #6562 \!strict
      Fix for unknown xml mime-type issue with IE.  Generic fix also fixes the file:// issue with xml files, but required minor UT change to not fail.  Test in question is a bit hokey to begin with, passing a fake xml DOM to the handler seems a bit .. odd (and doesn't really test much).  fixes #5388
      Tweak to message.  refs #6562 \!strict
      Fixing some broken links.  refs #6698
      Removing an obvious error.  A file had two copies of its contents in itself.  Reference patch:  dojo.tests.patch refs #6698
      Fix to ComboBox.  refs #6469
      Fixing issue with revert. \!strict   fixes #7145
      Fix for issue on Opera with Infinity as count value.  \!strict fixes #7269
      Making typeMap parser settable. \!strict  fixes  #7625
      Fixing _base/xhr.js for IE8.  refs #8137
      Committing in minor tweak to XML dom parser load path  refs #8137
      Committing in minor tweak to XML dom parser load path  \!strict refs #8137
      And now more patches for iframe and IE8.  Oy.  refs #8137
      Minor update to localise lookup of XML parsers by using _base private function.  refs #8140
      Fix for multivalued form field.  fixes #8212
      Adding in an abort linkup in ItemFile*Store to proxy the abort call. \!strict  refs #8315
      Tweak to testcase for abort to be nicer about timing (finish before abort is called), and only running in async cases, the browser. \!strict  refs #8315
      Reduce test group name size.  \!strict fixes #8372
      Minor cleanup in testcase code.  fixes #8403
      Minor fix to back. \!strict fixes #8431
      Update to IFWS.  \!strict fixes #8432
      Update to IFWS.  \!strict fixes #8432
      Tweak to fix case in null comparision.  fixes #8480
      Updated license text to clarify this and make it CLA/Dojo license compatible. \!strict fixes #8682
      Code reduction *and* fix.  \!strict fixes #9022
      Minor enhancement to ItemFile*Store + UT. \!strict  fixes #9077
      Making it possible to reset the store by setting either url or _jsonFileUrl.  \!strict refs #9303
      Patch to allow reset of data via the 'data' param.  \!strict refs #9322
      Minor tweak to close to avoid blowup during load if close is called.  \!strict refs #9322
      Adding in better custom type deserialization + UT.  \!strict (Also added in UT for #9322) refs #9332
      Adding in better custom type deserialization + UT.  Forgot a debug console.  Removed  \!strict refs #9332
      Tweak to sort function generator to improve performance.  refs #9099
      Missed a semicolon  refs #9099
      Minor fix to IFS with regards to hasAttribute, allows identifying empty multi-value attributes. \!strict refs #9379
      Minor feature addition. \!strict  fixes #8994
      Cleanup of styles dojo data classes. \!strict refs #10048
      Fixing syntax errors in test files.  Improperly escaped content in attributes.  refs #10063
      Updating eol in dojo.  \!strict refs #10405
      Committing in minor tweak. fixes #10474 \!strict
      Update item lookup in reference resolution case with auto-assigned identities.  \!strict refs #10583
      Fixing EOL style on a bunch of files to be 'native'.  refs #11151
      More EOLs refs #11151

Karl Tiedt (3):
      fixes #4516
      fixes #4818
      fixes #5464 - updating dojo.isDescendant to return false on error !strict

Kenneth G. Franqueiro (2):
      Committing backports of IE9/FF4 fixes to 1.5 branch, under SitePen CCLA. Fixes #14199 !strict
      Fix dojo.getComputedStyle code path for IE9; fixes #14103 on 1.5 branch !strict

Kris Zyp (17):
      Throw an exception when a DOM node is encountered in JSON serialization (fixes #6903)
      Throw an exception when a DOM node is encountered in JSON serialization (fixes #6903)
      Allows successful no-content responses (like status 204) to be successful with the JSON content handler. fixes #6601
      fixes #6765
      dojo.Deferred rewrite, refs #10814 !strict
      Fixes dojo.when handling of falsy values, refs #10814 !strict
      Make addCallback, addErrback, and addBoth only visible on Deferred (not promises), refs #10814 !strict
      Provides the fired property for backwards compatibility
      Fixes console target |this| on invocation, refs #10814 !strict
      Addition of dojo.Stateful for get/set/watch objects, refs #10839
      Make instances of dojo.Deferred each be an instanceof dojo.Deferred, and same for dojo.DeferredList.
      Support hash based sets, refs #10839 !strict
      Treat cancellations as unlogged errors, provide results property, refs #10814 !strict
      Fixes |this| for callbacks, refs #10839
      Adds error as results property, refs #10814, !strict
      Fixes global leak of fired/results, refs #10841, and adds missing require statement, fixes #11081 !strict
      Fixes issue with wrong |this| for cancellation of promises, refs #10814 !strict

Mark Wubben (1):
      Fixes issue with `new Deferred(function(dfd){ }).promise.cancel()` not passing the deferred instance to the canceller. Fixed `new Deferred().then().cancel()` resulting in a "deferred has already been resolved" error.

Mike Wilcox (13):
      Refs #7131
      Fixes #6588
      Refs #6588 !strict Removing some stray comments
      Fixes #7193 was not escaping a string in a certain condition when it was within an object !strict
      Refs #6588 !strict - Added check for previous version of Firebug. Should work well now with Firebug 1 and 1.2 - 1.1 is a bad and hard to find coe base, and should probably just not be supported./nAlso did a serious of minor tweaks to make this a little more jslint friendly. A little.
      Fixes #7534 !strict Added isAIR param so Lite won't launch in an AIR app
      Fixes #8601, #8602 - added trace() and count() methods. Doing an additional check for Mozilla's firebug lite. !strict
      Refs #9744 - Updated Firebug's browser sniffing. \!strict
      Refs #9744 - Changed the IE closure fix to a non-string closure \!strict
      Refs #9744 - removed the new keyword as per James' suggestion \!strict
      Fixes #9752 applying memory leak patch with changes as per James
      Refs #9752 applying memory leak patch with changes as per James - sending again, I think I forgot to save before the last commit.
      Fixes #9902 - Added Chrome 3 as a browser to not show Firebug Lite. \!strict

Nathan Toone (20):
      Fixes #6704: make internal call to _setValueOrValues instead of the API call setValues - to simplify extension of the class. !strict
      Refs #6921 - updated patch by haysmark to fix FF3 and FF2 issues.  See ticket for details !strict
      Fixes #7146 - add a definition for a data attribute so that the parser will pass it in !strict
      Fixes #7324 - allow array-based items to also have array-based sub-items !strict
      Fixes #7326 - account for padding and margin on the body when calculating position for onFirstMove
      Refs #7461 - document default value for position !strict
      Fixes #7642 - make sure consoleBody exists before trying to clear it !strict
      Refs #7865 - pass error through to callback !strict
      Fixes #7090 - add in patch that allows a mixin to be passed to dojo.parser.instantiate
      Refs #7090 - minor fix to when you specify types, need to make our value "look" like a namedItem
      Refs #8755 - add in optional flag for not starting widgets when calling parser
      Refs #9114 - add test cases which demonstrate breakage caused by changeset [17266]
      Quick-and-dirty fix to get things like the grid working in IE again short-term.  This fixes issues in IE caused by [17266] - and may not by the correct fix, but it prevents crashes. Refs #9114 !strict
      Refs #9131 - remove extraneous boolean checks for dojo.isString()
      Refs #9131 - remove extraneous boolean checks for dojo.isString()
      Fixes #9348 - fix the startup order of templated widgets
      Refs #9441 - add in formatting support extensions with respect to timezones
      Fixes #9618 - add in option for silent failing (defaults to false) to suppress the error message that gets consoled out on transport error !strict
      Fixes #9618 - add in test case (thanks Jared) and fixed constructor problem. !strict
      Refs #9862 - add test case for mixing in new functions to an instance of a class - this test passes with revision [20154] (though the mutatedMethods test breaks in that revision) and breaks in [20155] (which fixes the mutatedMethods test) !strict

Neil Roberts (29):
      Refs #4612. Lines can't start with keywords
      Refs #4820. Updated examples and descriptions that use code to work with tomorrow's API tool update.
      Refs #4915. Might as well clean up comments as well
      Refs #4820. Update documentation
      Refs #4820. Cleaning up documentation
      Refs #4820. Documentation updates
      Refs #4820. Documentation updates
      Fixes #5174. Remove 100 multiplication from gotoPercent
      Refs #4820
      Refs #5602. Cleaning up documentation !strict
      Fixes #2172. Edited summary lines that started with a parameter name.
      Refs #5602. Update pseudo-classes !strict
      Refs #5962. Mark code as code !strict
      Refs #5962. Revert marking code as code !strict
      Refs #5962. Markdown any problematic HTML !strict
      Refs #5962 and r12790. Update markdown and pseudo-args
      Refs #5962 and r12821. Instance variables won't show up in docs if there are no instance variables. !strict
      Refs #5962 and r12824. Properly type the browser detection variables !strict
      Refs #5962 and r12827. Make sure that nested lists are properly nested !strict
      Refs #5962. Remove parameters, they confuse people !strict
      Refs #5962. Update djConfig params
      Refs #6166. Somehow the window/document check didn't make it in !strict
      Refs #6166. Don't duplicate dojo.cookie !strict
      Refs #7250. Ensure that the first callback in the Deferred chain for an XHR operation (which processes the returned XHR data) prevents the received argument from being passed to the next callback.
      Refs #7461. Clean up returns statements !strict
      Refs #8606. Missing common in pseudo-code
      Refs #8606. Show callback functions as functions in documentation (with actual arguments) !strict
      Refs #8606. Show callback functions as functions in documentation (with actual arguments) !strict
      Refs #8606. Fix problem with pseudo-docs !strict

Nicola Rizzo (5):
      fixes #9440 - added a control on the counter before stopping the timer !strict
      fixes #6399 - added a new dojo.config property (transparentColor) to set a default
      typo. refs #8606 !strict
      fixes #9764 based on the sschuster's patch (thanks!), added a _getStep function to fix onBegin, onPlay and onPause too !strict
      small code reductions, fixes #10524 !strict

Nikolai Onken (1):
      fixed font-size for ie6/7, fixes #6755

Owen Williams (4):
      fixes #2629  -- ads prettyPrint option to dojo.toJson, added JSON tests to harness
      fixes #3227 Add dojo.toggleClass method.  See bug for details.
      fixes #3225  adding quick check so we don't perturb class on removeClass unless class was found
      refs #3151

Peter E Higgins (83):
      adding default list of dnd-based classes and make tests follow along.
      reverting parser.js changes from chengeset 9557.  broke: TabContainer,
      fixes #3405 - callback goes unused in example in documentation.
      fixes #2103 - by documenting the easing object some, and adding more inline docs for core fx + dojo.fx
      refs #4611 - small doc fixes to loader.js
      refs #4611 - cannot have (lowercased) reserved words as the first word
      refs #4661 - minor doc changes in html.js (can't start a line with 'returns') and adds in more verbose "description" vs "summary" in a few places.
      refs #5150 - add common blank.gif file to dojo base
      refs #982 - package level sup docs for dojo
      fixes #4969 - went with dojo.cookie.useObject (get/setObject combination? - contemplated asObject, and others. change if
      refs #5611 - ugly (safe) patch moves closure up to top. precedes coming patch to do other cleanups, that otherwise
      refs #3961 - tiny tiny cleanup to parser, and note on usage of dojo/method dojo/event
      fixes #5460 - makes public method fire() private (_fire) because you should never ever call it.
      fixes #5252 - sanity check on _started for _Widget
      refs #5866 - the onEnd tests for chain/combine work, but beforeBegin and onPlay never fire.
      refs #5539 - it doesn't actually appear to be broken (the ticket claim) though new unit test
      refs #5962 - show query().style() example in dojo.style() and document common pitfall.
      refs #3121 - who knew you could djConfig = { popup:true } ? this was broken in rc1, adding a test case to check for
      refs #5962
      refs #6021 - typo
      monster commit: refs #6963 - API change with deprecation. dojox.fx.easing is now called dojo.fx.easing, and all projects in dojox using
      fixes #7063 - add an .at() method to NodeList to select small subsets of NodeList's as a new NodeList - cla on file for Schontzler, David ... thanks!
      fixes #7406 - whitespace and duplicate variable declaration in html._ContentSetter ... thanks bitranch!
      refs #7461 - update fx.easing inline docs and style adherance
      refs #7553 - adds unit tests for dojo.fx.easing sanity checks.
      refs #7461 - updates inline docs for NodeList-html
      refs #7553 - make core html unit tests part of the test suite
      fixes #7604 - can't rely on djConfig being global, need to use dojo.config - which handles scopeName, and djConfig="" on the <script> tag, and conversion from a global djConfig
      refs #7461 - add example and words to animateProperty docs.
      fixes #6291 - adding package-level docs and fixing syntax of dojo.behavior API docs.
      refs #7671 - updating the format for inline docs in requireLocalization ... _browse test accurately find the params and
      refs #7461 - djConfig.require: Array was not documented.
      refs #7461 - commas are being evasive today
      refs #7461 - summary wants a :
      small doc update refs #7461
      refs #7461 - thanks neonstalwart for the doc patch/cleanup
      commiting additional fx unit tests to illustrate failure in Animation when calling stop() before a play() delay has started.
      fixes #8022, but went with .prototype instead of dojo.declare and fixes #7858 based on a [ccla][patch] from JavaStreet (bbyron). Thanks!
      fixes #8180 - m is undefined, so don't pass that. best bet, pass the index of the mixin that is missing. more helpful than #undefined. thanks dylanks.
      refs #6490 - unit test for combine chained and chained combined animations.
      i think this refs #7461 - a [meta] doc/style cleanup - but cannot reach trac to look up. adding example: tags and small style cleanups.
      fixes #8319 - reducing the mixing, caveat noted in ticket.
      refs #3961 - base size reductions, refs #7461 - adding many examples
      refs #7461 - commiting unsaved changes from previous commit. more examples.
      refs #7390 - giving vars breathing room, spacing var lists, etc. non-changing.
      refs #7975 and refs #7802 - adding dojo.destroy as an alias to dojo._destroyElement (leglizing the usage), and adds simple version of dojo.create following the API:
      refs #7802 - update unit test to not destroy firebuglite if in page. something about setting style:"" as an attr is wrong, so disabling that test for now and will investigate in attr itself as a separate thing. passes opera (though isFunction failed, investgate that too), and ie6/7
      refs #7975 - deprecating use of dojo._destroyElement internally in Dojo Core
      refs #7802 and refs #7461 - adding better docs for dojo.create (noting about alias to dojo.doc.createElement) and adding other examples.
      fixes #8317 - moving Toggler out into own file for 1:1 mapping and size considerations (though a require() is issued in fx.js for backward compatibility, it may be wrapped in a deprecated build tag as per #8271)
      refs #7975 - update to use dojo.destroy
      fixes #8298 - minor doc notation.
      refs #8353 - update things to use 2009 as end (c) date. Also seems a small stray comma fix snuck into this checkin.
      refs #7461 - fixing inline docs for nodelist (slightly). adopt and place pass through position to place(), which introduced
      fixes #8415 - adding .onchange support for nodeList. !strict
      refs #7461 - ducks quack (though are arguably quick, too).
      taking attr setting out of create. is redundant. these all technically beat enhancement cutoff, begrudgingly.
      fixes #8952 - adding foundation set/getObject unit tests.
      fixes #9007 - trunk is reopen "officially". 1.3 branch in place.
      refs #8606, #9063 - cleanup easing inline docs \!strict
      fun patch as mentioned onlist, notes:
      refs #4824 - officially exposing dojo.Animation and leaving an _Animation stub \!strict
      refs #9143 - doc and style updates for declare.js - no code changes, just part of the patch. \!strict
      refs #9143 - adding unit tests for fx showing inherited in a callback (two ways)
      fixes #8615 - adding unit tests to check the status of combined and chained animations onend
      refs #8606 - allow doc parsing to continue. fixing syntax error in pseudo-block
      refs #8612 - adding in a failing unit test that works before [17640]
      fixes #9435 - small whitespace issue (space after pipe needs to be a tab) - tested in _browse.php
      englishifying and expanding the docs for coords and position a tiny bit. refs #9648
      tiny nit refs #9648
      refs #8606 - expand dojo.hitch docs just a bit
      refs #8606 - making the pseudo-docs parse properly \!strict
      fixes #9503 - api docs should pickup defaults from members. added wordage indicating members are mixed over from constructor, expanded description some. moved inline docs around a tad, and added some TBD fixme's
      fixes #9806 - adding in foreach of removeAttr and unit tests. Added notes to campus docs
      fixes #8707 - exposing _contentHandlers, adding docs, keeping alias. added to wiki docs
      refs #8606 - adding tabs after | in examples so they markdown properly. expanding docs a tad \!strict
      refs #9984 - giving root object a summary. leaving bug open, the kwArgs are properly pseudo-doc'd so this might be a parser bug. not mixing standard IoArgs into a fake declare
      refs #8606 - expand docs and inline examples. adjust alias (though should be moot now) !strict in html.js
      fixes #10040 - expand inline docs for core fx methods
      refs #8606 - update inline dojo.cache docs
      refs #10704 - updating unit test to pass at() when firebug present, who is now adding a <div> to <html> for firebugConsole. scope the selection a tad
      fixes #10036 - make .at() allow negative indices with test and extended docs
      fixes #8721 - introduce auto:true support to NodeList-fx animations to provide a way to automatically play the animation and return the nodelist for further chaining.

Revin Guillen (1):
      Fix a regression in which addClass and toggleClass can no longer handle and undefined classStr parameter. (refs #11146) !strict

Robert Coup (3):
      Fixes #3801
      Revert [9708] and make dojo.requireLocalization smarter about local vs xd when we're in a cross-domain scenario. Break out check for xd or local from _loadPath to support it. Refs #3801
      Add a warning instead of throwing an obscure exception if dojo.back.init() hasn't been called. Fixes #3861.

Sam Foster (11):
      Fixes #7157, corrects the docs in dojo.place !strict
      removed unused parseOnLoad alias for parseContent, fixed error in the fallback error handler which was refering to this.domNode. refs #5647
      Refs #5647
      Refs #7560
      doc patch on onFirstMove to call out the px unit assumption; fixes #8874
      new dojo.html._setNodeContent implementation to leverage dojo.place and dojo._toDom, plus nodelist, mixed content unit tests. Fixes #8224
      Remove defunct shouldEmptyFirst 3rd param to dojo.html._setNodeContent Refs #8224
      DocumentFragments can just pass thru to dojo.place, cleaning up enumerable handling. Refs #8224
      Added test for :checked psuedo as applied to <option selected> (per CSS3 spec). Refs #5179
      Adding missing 'throw'. Fixes #10286
      Moving throw to set(), where having no .node property is a show-stopper. Instantiating _ContentSetter without a .node is valid as reuse of the setter is encouraged. Fixes #10286

Scott J. Miles (79):
      First draft of new connect/event stuff. Refs #2641
      Event.js: all browsers: set keyCode to 0 when charCode has a value, norm CTRL-BREAK to CTRL-c where possible, IE: use native keypress for ESC and ENTER
      Repair test by using dojo.marginBox to measure height. Refs #2710.
      Rearrange dojo.style so it doesn't call getComputedStyle unless needed. Refs #2686.
      Help fx.html work in IE: ensure integer endHeight, set overflow: hidden to allow resizing test node. Refs #2848.
      Remove vestigal argument to "dispatcher". Refs #2641.
      Check for built-in constructors that do not expose "apply". Refs #2691.
      Make sure stealth onkeydown listeners are disconnected with their progenitors. Fixes last known missing requirement for event.js. Refs #2641
      foo should be "foo", my bad. Refs #2848.
      Add another trap for read-only keyCode on IE, refs #2882.
      dojo.event.addListener was returning the wrong function, refs #2894.
      Another round of revisions to connect and event. Refs #2641.
      Fine tuning in connect. Redo IE keyboard magic in event.js so it actually works. Update docs. Refs #2641.
      Fix bug (x2) setting stealthKeyDown flag. Thanks Doug. Refs #2641.
      On Mozilla, printable chars have a 0 keyCode, trying to simulate that on IE clobbers regular key handling. Refs #2958.
      Fix test that was using the console instead of the test object. Refs #2641.
      Error trap in 'inherited' was in the wrong place. Refs #2692.
      Special event handling for CTRL-ENTER on IE. Refs #2886.
      More repairs due to side-effects of trying to clear keyCode as on Mozilla. Just eliminating that part of the normalization for now, it's not necessary. Refs #2886.
      Tweaking the box getters and setters to deal with scrollbars, change boxMode to boxModel, assume border-box on IE5.5. Refs #2686.
      Make _getMarginBox return proper left/top. Refs #2686.
      Fix mixin specifity issue. Additional general optimizations. Refs #3030.
      Fix typos. Refs #3030.
      Make _getMarginBox left/top always be at least 0 (vs. NaN). Refs #2686.
      Messed up and forgot to commit updated test. Refs #3110, #3030.
      Change some "this" to "dojo". Library functions are generally callable in any scope. Refs #2686.
      Add an exception for a bad hitch, refs #2609.
      Trap extra calls to Object constructor. Refs #3030.
      Inline docs for subscribe/unsubscribe/publish, refs #3161.
      Minor comment edit on subscribe, refs #3161.
      Update test for new unsubscribe syntax, refs #3154.
      Privatize add/removeListener, client code should just use connect/disconnect. Refs #2934.
      Fix _toStyleValue caching, refs #3295.
      Non-ideally, remove assertions that are not be on WebKit, allow tests to pass. Refs #3198.
      Safari: add a null test and return a golem object from getComputedStyle if necessary. Refs #2686.
      Add DnD test for parent_constraints with margins. Refs #3320.
      Forgot to account for border when switch between client and offset sizing. Refs #3297.
      Add overflow: hidden to parent in DnD test to avoid confusion about the results. Refs #3320.
      Unify connect and event listener interfaces. Change disconnect package to hold a direct reference to a listener. Introduce IE6 leak protection for custom (non-DOM) events on DOM-nodes. Refs #3391.
      Repair blunder in connect, now connect target should correctly return it's value. Had to effect the change in the leak-proof listener also. Refs #3503.
      Refactoring of html.js. Refs #3552.
      Restrict Safari margin hack to nodes that are not position: absolute, allowing more tests to pass, refs #3515.
      Change addOnUnload events to fire at window.onbeforeunload (except on IE), refs #3567.
      Numerous additional tests for html.js courtesy of Steve Orvell (TurboAjax Group). Refs #3566.
      Fix test names. Refs #3566.
      Prevent _getMarginBox and _sumAncestorProperties from calling getComputedStyle on non style-bearing objects (i.e. document). Refs #3620.
      Comments hopefully clarified, refs #2556.
      Implement mixin semantic as an inheritance tree to support robust "inherited", refs #4068. Support improved initializer syntax, including deprecation warning for older style, refs #4069.
      Reduce footprint of boilerplate constructor by simplifying closure, refs #4068.
      Additional (crufty) filter for native toString implementation in dojo.mixin, refs #4071.
      Don't select a prototype tree for 'inherited' that doesn't contain the target method at all, refs #4068.
      Less zealousness when clobbering keyCodes in preventDefault for IE, refs #4064.
      Monster commit to update uses of declare to new cleaner syntax, refs #4069. Changes are 99% syntactic only. If you have a problem, let me know and I will help you fix it asap.
      Fix files that got messed up when merging changes, refs #4069.
      Make sumAncestorProperties do what it says it will do (+test), fixes #1965.
      Make toPixelValue translate 'medium' to 4, hack refs #3874.[[BR]]
      [[BR]]Aggressive leak protection on IE7 (which is more broken than we had hoped), fixes #4141.[[BR]]
      Comment out unused dependency on PEAR, fixes #4158.
      Prevent duplicating 'preamble' in intermediate delegates, fixes #4243.
      Call "postscript" method after construction, refs #4243.
      Add 'extend' method to declared classes (not instances), fixes #4348.
      Repair _watchInFlight loop, fixes #4386.
      Style faux pas, refs #4386.
      "." becomes "+", fixes #4449.
      Ensure non-null 'props' before dereferencing, fixes #4532.
      No reason to put intermediate delegates into any namespace. Prefer simply removing this code. Refs #4839.
      Fixes #4908 (thanks Bill). !strict because the pre-commit hook is confused about return types.
      Fixes #6050, make sure to clear _stealthKeydownHandle, so key listener can be reconnected later.[[br]]
      Must pass calling context on event handler through to event fixer to locate the correct currentTarget and window.event. Fixes #6069. Fixes #6050. !strict
      Fixes #5372. !strict
      Do not assume baseclass implements _findMixin. Refs #6696. !strict
      Extend "inherited" method in declare to use memoized prototype information when available instead of searching. Should improve speed of "inherited" and also fixes #6846. !strict
      Uncomment nulling of preamble in delegate classes, viz Nathan's patch, fixes #6846. !strict
      Detect 'border-box' for 'input' tags of type 'button'. Fixes #5518. !strict
      Return {} if input to getComputedStyle is not Element. Fixes #6657. !strict
      On Safari, fix "button" edge case in "setMarginBox". Fixes #7148. See also #5518. !strict
      Fix variable name typo, refs #7148. !strict
      Iterate over cloned listener arrays in case the original arrays are modified during event processing. Fixes #6165. !strict
      Manual Array copy in connect.js under Rhino. Fixes #7523. !strict

Simon Bates (4):
      Added new API for getting and setting HTML attributes: dojo.hasAttr(), dojo.attr(), and dojo.removeAttr(). Fixes #5055. !strict (errors outside of my changes)
      Minor doc fix. Refs #5055. !strict
      Fixed up the _base/html.html unit test file to pass on Opera 9. I disabled some tests that cannot pass (tabindex="-1"). Refs #5055.
      Small code simplification for dojo.hasAttr(). Refs #5055. !strict

Tom Trenka (15):
      Refs #2597.  Start of a default stylesheet.  Includes some (but not all!) of YUI Reset, and defines a baseline for content-based materials.
      Refs #2597.  (Hopefully) proper attribution made WRT YUI Reset; added table definitions and tests.
      Refs #2597.  Uses @import to pull in YUI reset, to avoid IP issues.
      Refs #2597, adds rule for ordered lists.
      Refs #2597.  Moved test to new test directory.
      Refs #2597.  Changed font to Tundra default.
      Refs #2597.  Consolation of some rules.
      Fixes #3051.  Provides basic inline documentation as to the purpose of the stylesheet, and what it will do to a web page that uses it.
      Refs #4209.  Testing for the dojo stylesheet change with the addition of the dojoTabular class for tables.
      Refs #6334.  Adds XML support to dojo.io.iframe; nasty but working hack on IE makes this go ok.
      Refs #6334; added a CDATA block to double check the test.
      Refs #6334.  Forgot to turn off the display of the iframe so I can inspect the contents.
      Refs #6334.  Updated documentation on dojo.io.iframe.send.
      Fix the iframe.__ioArgs documentation definition.  Refs #7496, !strict.
      Small change to __FormatOptions definition.  Refs #8606.

Torrey Rice (1):
      removing Myriad Pro from the CSS for various reasons, one being that there are a number of overlapping issues in Firefox. The other being that most people don't have this font anyways so we shouldn't really rely on it as our default font. fixes #4429

-----------------------------------------------------------------------

This annotated tag includes the following new commits:

       new  b23d8a6   Using r22487 to create a branch named 1.5.
       new  43584e0   Fixes #11510: fix issue with regexp looking like a comment for debugAtAllCosts removal.
       new  efba883   Fixes #11575 !strict.  Change dojo.version to 1.5.1 in 1.5 branch.
       new  13c5db4   Fix parser to obey _noScript flag in prototype again, so that dijit.Declaration works.   Also adding unit tests for dijit.Declaration.  Fixes #11586 !strict.
       new  59d48c7   Refs #8546, allow using protocol-relative URLs with xd loader (backport to 1.5 branch)
       new  8df0724   Remove unicode bidi control chars, 1.5 branch.  Fixes #10847 \!strict
       new  898f164   Kazakh translations Refs #11648
       new  aef4dae   patch from retchless \!strict fixes #11257
       new  6aceabc   Updated translations. Fixes #11648
       new  a8b9c14   Applying Bryan's patch to fix am/pm overrides.  Fixes #12155
       new  ccb7a51   translation corrections.  Refs #12322
       new  950dfb3   Committing backports of IE9/FF4 fixes to 1.5 branch, under SitePen CCLA. Fixes #14199 !strict
       new  3e66f1a   Fix dojo.getComputedStyle code path for IE9; fixes #14103 on 1.5 branch !strict
       new  bf99789   IE10 even in quirks mode does not support filters for opacity.  Fixes #16117 !strict.  Backport to 1.5.
       new  4b50371   Use per element determination of children vs childNodes, fixes #12488 !strict
       new  1708abc   Updating metadata for 1.5.3

The 16 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails.  The revisions
listed as "adds" were already present in the repository and have only
been added to this reference.


-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-javascript/dojo.git



More information about the Pkg-javascript-commits mailing list