[Pkg-javascript-commits] [uglifyjs] annotated tag upstream/1.3.5 created (now e1df2df)
Jonas Smedegaard
js at moszumanska.debian.org
Tue Apr 29 23:32:53 UTC 2014
This is an automated email from the git hooks/post-receive script.
js pushed a change to annotated tag upstream/1.3.5
in repository uglifyjs.
at e1df2df (tag)
tagging 222825b87ed3e9fda3960e2cf1bf174c4e7a4e54 (commit)
replaces upstream/1.3.4
tagged by Jonas Smedegaard
on Wed Apr 30 01:22:03 2014 +0200
- Log -----------------------------------------------------------------
Upstream version 1.3.5
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=
=gouD
-----END PGP SIGNATURE-----
Aaron Blohowiak (1):
make the module exports a convenience function so minifying src with the defaults is just a one-lininer: min = require('uglify-js')(src);
Andrey Tarantsov (1):
Fix broken calls of complex binary expressions [#471]
Cedric Vivier (1):
Extend name mangling from base54 to base64 after first character. Provides better compression on libraries with a large number of variables.
Chris Hoffman (1):
Make the package.json work when npm linking again
Eirikur Nilsson (1):
Added a failing test for with statements.
Emil Hesslow (12):
Added an easy way to create unittests and created units test for 16 of the fixed issues on github
Add support for compressing 'a + "str1" + "str2" + b' to 'a+"str1str2"+b'
compress assignments more
compress assignments more
Be even better with assignments
Be even better with assignments
Treat the first statement in a for-loop the same as a normal statement. So do seq with previous statements. Put together var and const declarations. And if it is empty move the last statement there
Add some more types to find_first_execute. Gives a few bytes on jquery and processingjs
Add a decompress step first so we don't have to add a lot of code to handle special cases around sequences and first statement in for-loops.
Move i++, ++i, i+=1 forward into the next statement if possible
Fix issue 48 where 0 and "" was treated as same and wrongful optimizations where done
- Removed --extra because of bugs in the optimizations. After the bugs are fixed I will move them into --unsafe so they get some testing.
Fintan Fairmichael (1):
Fixes printing printing small floating point numbers.
Isaac Z. Schlueter (1):
Remove whitespace in #! line
Jann Horn (2):
compile `a=a+b` to `a+=b` and so on
put the assignment types into an array literal instead of stringsplit
Jason Davies (1):
Fix sparse array literals e.g. [,]
Jez Ng (1):
Allow defines even when not mangling.
Jonas Smedegaard (3):
Merge branch 'upstream-experimental' into upstream
Strip .gitignore file.
Imported Upstream version 1.3.5
Joseph Gentle (1):
Added null as a constant.
Mal Graty (6):
Improved strict mode support
Don't bother save/restoring in_directives
Apply ascii only option to regexps
Squashable directives
Adding tests for squeezing directives
Fix make for newer versions of node
Marcelo Jorge Vieira (1):
Imported Upstream version 1.3.3
Marijn Haverbeke (1):
add an ascii_only option to gen_code
Mathias Bynens (2):
Only a subset of whitespace characters is problematic after `</script`, so use `[\t\n\f\r ]` instead of `\s`.
`</script/` is problematic, too.
Matt Glazar (4):
More sane unit testing script
Merge commit '9472f4ff19e0daf3adb7fdb0304ab77f19d3cc52'
Compress ifs further
Failing test cases for name mangling
Michael Ficarra (2):
version test for node was missing a leading `v` and was not compatible
fixed issue #112: `_ == true` -> `_ == 1` and similar
Mickaël Delahaye (1):
Add syntax error on newline after 'throw'
Mihai Bazon (440):
git init (moved from a mercurial repo, but I don't care about previous history)
minor
fixed some issues with code generation; it's now hopefully complete.
added process_ast which mangles variable names and will eventually apply
resolve simple binary ops on constants
mangling/compression finally works fine (compress itself, run the compressed
fix parsing problem (only allow minus as part of the number when it's the first character)
allow function(){return} (no semicolon after return) when not in strict-semicolons
"new Foo;" should still insert an empty argument list
added "with" and "atom" generators for codegen
handle return with no arguments in codegen
fixed parsing conditionals (thanks Marijn!)
support "with" and "atom" in process_ast
minor
disorganized tests
always add space between consecutive [+-] operators (handles "i + ++j").
add parens around called expression if not "name" or "dot" (should cover all cases)
display empty blocks as "{}"
minor
Better fix from the master programmer for some issues that I identified:
better fix for "{return}" case (http://marijn.haverbeke.nl/git/?p=parse-js;a=blobdiff;f=src/parse.lisp;h=da8409d1e7078dea43c5c5de7ce5afcf1caf6562;hp=80438c39afad5ca5d5fc3fca35a13d3daae63de3;hb=732969c989144e56a74b51dea6b9e227a7e9d031;hpb=2e09d491179be4a838e9435cdacfd7f1e84e9646)
fix for parsing ".1" (discard the dot from punc. chars)
minor
minor
minor style changes
rewrote proper mangler/compressor (in process.js); revised code generator.
some reorganization, some fixes (moved gen_code to process.js, handle
minor
Fixed bug that showed up when processing jQuery: I was using the "token in
lots of other fixes while testing on jQuery:
another fix: don't shadow a name that is referenced but not defined -- it's possibly a global defined elsewhere
dropping spurious brackets got back, but this time it's safer
further optimizations for IFs:
more optimization for IF statements
comments + small cleanups
minor
added some documentation
README.org for github
symlink not good enough for github.
README for github, take 3
fixed doc rendering for github
minor doc changes (link to the Google Group)
links
links
squeeze a few more bytes: if (!foo) bar(); else baz(); ==> foo ? baz() : bar();
minor: transform !foo?bar():baz() ==> foo?baz():bar() for all conditionals (16 less bytes on DL and 3 on jQ :-p)
Fix for: http://github.com/mishoo/UglifyJS/issues#issue/2 --
Allow unescaped slash character in literal regexps, when inside a character class.
we must add semicolon after do{...}while()
newline after return shouldn't cause the next token to become the returned value:
maybe_before_semicolon no longer in use
add parens to "seq" in conditionals; fixed typo that caused this to break in
support for '2e-3' (http://github.com/mishoo/UglifyJS/issues#issue/4)
newline normalization (http://github.com/mishoo/UglifyJS/issues#issue/7)
small fix (content of line comment tokens had the last char cut)
Multiline comments that contain a newline will trigger semicolon insertion
when not in "strict semicolons", allow trailing comma in hash and array literals
add parens around "assign" when it's used as condition in the ?/: operator.
add parens around "seq" when used as constructor.
fixed again http://github.com/mishoo/UglifyJS/issues/issue/11/#comment_361674
`new (X ? A : B)` → `new X?A:B` --- now fixed.
Fixed parsing [^[ ... ] in regexp.
adding space before the colon in literal objects is now optional
minor cleanup
some cleanups
removed old/unused files
no more JSON.stringify to output strings
small script to call the uglifier/beautifier
add parens around "seq" when used as condition in the ?/: operator
minor improvement: if (!foo) bar(); else baz(); ==> if (foo) baz(); else bar();
minor: don't add parens around "binary" when used in combination with an operator of the same precedence
reverted previous commit [ a-(b-c) was translated to a-b-c which is wrong ]
minor
block brackets are required in try/catch/finally
small cleanup
some cleanups regarding "block" nodes -- and gained a few more bytes by properly adding the brackets *only* when really needed.
Fixed bug when adding blocks to THEN clause:
fix: don't add a mangled name to anonymous functions
minor optimizations/experiments:
minor fix gains another few hundred bytes on DL, Ymacs and jQuery.
default make_seqs => true
update docs
fixed links
no need to escape slashes in regexp, since we store the backslash as well
fixed parsing for x++ / 5
minor
escape \r and \f when in output strings
parenthesize "dot" in "new" calls; sometimes this can add unnecessary
back to using JSON.stringify to output strings (still the safest variant)
support "const"
fixed parsing 0xC1BDCEEE
fixed issue with mangling: http://github.com/mishoo/UglifyJS/issues#issue/17
added option to mangle toplevel scope
minor improvements:
Handle EOF when reading blocks.
minor fix for numbers
fixed again parsing unterminated blocks (handle EOF)
don't enforce for/do/while/switch in label_statement
in fact, do check for do/for/while/switch after label, but only in "strict_mode"
croak on try{} without catch{} and finally{}
added --ast
when --ast is specified we should still go through all the requested tree processing
"case" and "default" nodes no longer appear in the tree; instead, since they
removed unused code
added --overwrite to overwrite the original file
Merge branch 'master' into macros
remove final "break" statement in switch (only when unlabelled)
some doc updates
allow regexp immediately after else
allow a literal regexp to start a statement
looks like comma-expressions should be explicitly forbidden in expressions of the "conditional" (?) operator.
Accept operator &=
disallow assignment to 'this'
NaN is not a keyword
when parsing a function, reset in_loop to disallow break/continue
mark 'i' as referenced in the current scope on "for (i in foo)"
tighten try/catch/finally blocks
ability to remove some dead code, and warn about it (ON by default):
updated doc
update doc
fix for http://github.com/mishoo/UglifyJS/issues#issue/25
some cleanups/fixes that were figure out in the Lisp version [1]. Big speed improvement.
when the condition in a conditional is a conditional, should add the parens
don't quote object keys when they are integers
cut off a few more bytes by being a bit smarter about adding parens around function literals
more cases to handle before discarding parens around function literals
skip literal functions when detecting calls in "new" expression
handle <<, >> and >>> when resolving binary operations
should be safe not to force numbers in evaluating <<, >>, >>>
more accurately support numbers as keys in literal objects
fixed regression (we always quoted object props)
fixed another regression (we always added parens in unary-prefix/postfix)
discarded other unnecessary parens -- before, !!b used to turn into !(!b)
convert string properties to numbers where possible in literal objects
reverted previous commit (not such a good idea, as it changes semantics)
fix parsing for 1e+2
in Node-0.3 "sys" has been renamed "util"
minor
Merge branch 'master' of http://github.com/michaelficarra/UglifyJS
make sure *this* is quoted
Merge branch 'master' of http://github.com/satyr/UglifyJS
sys = require("util") for node 0.3+
some cleanup
squeeze a few more bytes with an apparently safe transformation:
Small note: we beat GCL on jQuery ;-) Should update the stats in the documentation.
small change in handling the first comment (only keep the first X comments
small change in the way the lib is loaded (officially named "uglify-js";
Merge branch 'assignment' of http://github.com/WizKid/UglifyJS into wizkid-master
whatever I'm commiting right now should just be a merge operation
Merge https://github.com/WizKid/UglifyJS into wizkid-master
renamed uglify-js.js to index.js since it appears to be too complicated to
Merge commit '38f691e'
minor optimization for if (foo) { ... return }
Merge branch 'wizkid-master'
fixes issue #43 (restore regex_allowed after skipping a comment)
minor
removed unused code
added --extra option, which will enable optimizations that we're not yet
fixed issue #50 and added test case. also:
Merge remote branch 'wizkid/master'
oops, fixed regression (var joining is safe, should be enabled by default)
true => !0, false => !1 -- seems safe; this cuts 1K on DynarchLIB
be smarter about negation:
fixed issue #53: x = (y, z) should not discard the parens
foo.toString() ==> foo.""
removed useless check
disabled foo.toString() ==> foo+""; it's still available if you pass
minor (got rid of try/finally)
fix for issue #58
fixed issue #57 -- add block brackets around do/while in code like this:
fixed issue #59 ("!--x" incorrectly yield "--x")
fixed issue #61: f((a, b)) was incorrectly rendered f(a, b)
removed Node dependency from process.js (issue #60)
fixed issue #62: properly handle negation such that:
got rid of JSON.stringify for writing strings; having our own function now,
compress new Array(1, 2, 3) into [1,2,3]
fixed regression: new Array(x) should be left as is.
properly handle the case where Array exists in current scope:
updated expected test result (broken since we added support for quoting with
but then indeed, "new Array(x)" can be safely written "Array(x)" if we're
compress Array(1,2,3) into [1,2,3] too
updated doc to point out the potentially unsafe Array transformations
doc fix
added an array mapping utility that allows us to push stuff at the top
fixed issue #68: push function definitions to the top of the current scope
updated "expected" version for tests affected by the last change
fixed issue #69
fixed issue #70: for reasons unknown to humanity, you can have an
switched to the simplified BSD license
support parsing getter/setter declarations in object literals (patch by Pavel Podivilov)
support mangling/squeezing/gen_code for getter/setters in object literals
fixed comment
issue #71 -- increment line count in multiline comments
issue #72: (0).toString -> 0..toString
fixed parsing "5..toString()"
support [1,,3] (an array with 3 elements)
Added the ability to insert some newlines in the generated code
moved split_lines into the library code and exported it
reversed the previous patch; the ideea was good, but implementation seems to
fixed issue #77 -- endless recursion if the stream ends with ++/--
remove bogus ~= operator and extra %= mention
fixes #78 -- quote numeric property names when they're negative
don't call output.end() when we're writing to standard output
semicolon removal less aggresive:
Employ the full parser, rather than just the tokenizer, for splitting
read --max-line-len argument as an integer
fixes #83 -- show a cleaner error when the input file isn't found
fixes #84 -- safer dealing with boolean expressions,
refs #84 -- consider a = !b a boolean expression too
transform !(a?b:c) ==> !a?c:b
minor optimizations:
oops, fixed regression: !(a?b:c) can't convert to !a?c:b
fixed negating conditionals again :-\
fixes #86 -- dropped unsafe transformation
mostly cosmetic change
removed unused code
removed more unused code
fixes #65 -- moved the following transformations under --unsafe:
keep the first token in ast_walker, ast_add_scope and ast_mangle (useful for
more embed_tokens: do lambdas too
fixes #87 -- transform while(true){...} into for(;;){...}. Drop the block
but don't try to warn when the "unreachable code" is empty
resolved conflict
don't display warnings unless --verbose
fixes #94 -- drop names of inline functions when they aren't referenced anywhere
slightly better fix, in case your inline function is named toString,
Merge branch 'master' of https://github.com/cadorn/uglify-js into cadorn-master
fixes #92 -- added --reserved-names command line argument (pass one or more
updated doc of ast_mangle (there was a small API change)
issue #96 -- added parser test suite, thanks to Peter van der Zee
issue #96 (regular expression flags) -- accept any identifier character for
support unicode chars in identifiers
disabled some checks by default (they're enabled in strict_mode):
treat ZWSP (zero-width space, U+200B) as whitespace
issue #96 -- support UnicodeEscapeSequence in names
fixed some regexps
minor
allow new Date++
allow call expressions to be lvalues too
don't check whether the left side is assignable, unless strict_mode == true
support funny cases of for-in:
renamed variable "strict_mode" to "exigent_mode", to avoid confusion with
reverted 687382e4d2f55401483e6823cb814d7c9cab5dbc and updated testparser.js
fixes #99
warn when "conditional comments" are encountered
added --ascii option, useful to encode non-ASCII characters as \uXXXX.
updated gen_code API description
transform a["123"] into a[123] (fixes #107)
only add a second dot in forms like (5).toFixed() if the number is integer.
fixed parsing for stuff like 0x12.toFixed() (when we have 0x, the dot cannot
removed duplicated operator
fixes #117 -- don't parenthesize literal objects
add parens around object literal when used as first thing in a statement
fix more pathological cases of parens around literal object
again: it can be first thing in a statement when in unary-prefix too...
more parens around literal objects/functions when first thing in a statement
handle conditional operator too
NPM changes
comments (fixes #120)
Merge branch 'master' of https://github.com/chriskr/UglifyJS into chriskr-master
fixes issue #129 -- initializers in var/const cannot be seq-s (parens needed)
only defuns define a new name in the current scope; named lambdas seem
better fix for issue #128:
support "splice" node, which is like "block" but doesn't add the brackets in gen_code
added a sample on how to use the parser/walker to instrument some code
embed tokens in expr_atom and expression parsers too
add brackets around splice nodes when they're directly under "for",
Merged pull request #133 from aaronblohowiak/master.
Merged pull request #137 from tav/master.
can't optimize away do{ ... }while(false) - fixes #150
options.defines is optional..
Merge pull request #149 from killdream/master
simple "make" utility. ideas welcome.
v1.0.2 -- fixes #152
newline after continue/break ends the statement (fixes #154)
minor
mangled name of a function expression should not clash with a name in the
version 1.0.3
small cleanup
fixed mangling of lambda names (regressed as of ae88177)
version 1.0.4
fixed regression after ae88177fde6a7e54b047: mangling the name of function
give unary postfix ops the right precedence (fixes #173)
added --inline-script option to escape `</script`, by popular demand
updated doc
version 1.0.5
minor (`</script` needs escape only if followed by `>` or whitespace)
Merge pull request #177 from mathiasbynens/master
support deprecated OctalEscapeSequence in strings (fixes #178)
Merge pull request #179 from eirikurn/master
fixed compression of `with` statements (regressed at 9053435497e6ab53b100ead72aa399e48f3df8c3)
version 1.0.6
parenthesize seq values in object literals (fixes #180)
discard backslashed literal newlines in strings (fixes #183)
Added ast_lift_variables, which does a little more than its name says:
attempt to group assignments into the var declaration
minor
more solid check need of semicolons after the last statement of a block.
Easy yet safe optimization. Saves 119b on jQuery.
Updated readme. Follow @UglifyJS on Twitter. ;-)
no point to keep initialization values in var/const in unreachable code
mangle label names (fixes #151)
more fixes:
Merge pull request #199 from thorn0/patch-1
moved Array ==> [] transforms in squeeze-more (so they now apply on `--unsafe`)
Merge pull request #200 from johnjbarton/01a83b461f541898cb4b2f9ef033245c1a7ee18b
Merge pull request #202 from lautis/fix-broken-tests
Merge branch 'master' of github.com:mishoo/UglifyJS
updated doc
exit with code 1 in case of error (fixes #204)
version 1.0.7
fixed parsing in exigent_mode + embed_tokens (fixes #208)
merged cade4f252e5f4db715d9 from @schmerg, other minor doc updates
fixed codegen for `(function(){...}()) && bar()`.
Added an utility that allows walking depth-first easily.
fixed rendering of extremely "large" negative numbers (fixes #214)
err, fixed again negative numbers. (#214)
Merge pull request #220 from proxv/master
Merge pull request #216 from jasondavies/sparse-arrays
Merge pull request #230 from thejh/assign-to-modifyassign
support empty "then" IF branch in codegen
fix for parsing regexps like /=/ where statement is expected
Merge pull request #232 from proxv/master
add parens around literal object as statement (fixes #234)
Merge branch 'master' of github.com:mishoo/UglifyJS
fixed removal of `if (false) { ... }`
Merge pull request #235 from josephg/patch-1
Merge pull request #237 from proxv/parenthesize-no-prop-object-literals
Merge branch 'issue112' of https://github.com/michaelficarra/UglifyJS into michaelficarra
Merge branch 'master' of github.com:mishoo/UglifyJS
Merge pull request #172 from 5long/master
semicolons
Merge pull request #211 from lautis/unicode-whitespace
in fact, new Error().stack is enough to get a stacktrace
Merge branch 'master' of github.com:mishoo/UglifyJS
try negation in make_conditional (closes #227)
Merge branch 'encode-nulls' of https://github.com/lautis/UglifyJS into lautis
version 1.1
fixed version for NPM
Merge branch 'parenthesize-function-expressions' of https://github.com/proxv/UglifyJS into tmp
Merge pull request #238 from proxv/multiline-strings-semicolons
keep literal newlines in strings as \n
reverted last patch
fix parenthesization of anonymous function calls
better fix instead of f896f690500f11130435c495a7b1eaf9338c59dd
negating >= and <= operators is back to --unsafe (closes #239)
add final semicolon (closes #126)
version 1.1.1
use == instead of === in `member` (fixes #244)
escaped slash character in regexp
String(a) => a+"" (with --unsafe)
at most one ';' at EOF
some cleanup
Merge pull request #263 from ikokostya/object
Use a different scope for labels than for variables.
Merge pull request #258 from ikokostya/fix-error-pos
Merge pull request #248 from WolfgangKluge/endpos-in-tokens
fixes embed_tokens for functions (#260)
option to prevent mangling or dropping of function names (#267)
updated doc
version 1.2
made npm happy
don't walk label statements twice (closes #272)
version 1.2.1
walk label statement *after* mangling it (really fixes #272)
v1.2.2
support `debugger` statement (closes #278)
improved redo_if (closes #283)
v1.2.3
Merge pull request #291 from isaacs/patch-1
v1.2.4
Merge pull request #295 from WolfgangKluge/master
handle null constant (closes #299)
v1.2.5
avoid parens for getter/setters (closes #301)
avoid double-parens on calls on function expressions (fixes #313)
output literal TABs in strings (fixes #318)
fix parsing `x = {}/1` (closes #320)
Merge pull request #327 from RGustBardon/master
Merge pull request #328 from RGustBardon/fix-162-and-271
Merge pull request #329 from RGustBardon/implement-325
v1.2.6
support "use-strict" as a statement (refs #339)
Merge pull request #345 from mal/master
Merge pull request #355 from proxv/debugger-semicolon-insertion
Merge pull request #356 from proxv/fix-failing-test
Merge pull request #357 from proxv/handle-comment-newlines
Merge pull request #358 from proxv/package-improvements
Merge pull request #359 from proxv/fix-consolidator
Merge pull request #369 from mal/ascii-regex
Merge pull request #373 from mal/master
fix parsing 1e5..toFixed(5)
don't generate double-dot for 10000..toFixed(5)
Merge branch 'master' of github.com:mishoo/UglifyJS
avoid some double-parens
Merge pull request #342 from int3/master
Merge pull request #384 from RGustBardon/fix-383
Merge pull request #385 from RGustBardon/implement-368
Merge pull request #387 from rvanvelzen/master
when the statement of a for is not empty, check if it requires semicolon
Merge pull request #397 from rvanvelzen/optimize_gzip
Undo commit 133b36f54d0e8f78fc649e99983798bf05445b18
Merge branch 'fix-222-and-324' of https://github.com/RGustBardon/UglifyJS into RGustBardon-fix-222-and-324
propagate newline_before for comment tokens (fixes #398)
when --unsafe, replace "foo".toString() with "foo", rather than "foo"+""
Merge pull request #400 from proxv/override-digits-in-tests
Merge pull request #401 from proxv/ast-squeeze-call
v1.3.0
ast_mangle should, well, mangle by default. (fixes #404)
also don't mangle function names if !options.mangle (#404)
fix options for ast_mangle
Merge pull request #407 from RGustBardon/1e+21
Merge pull request #408 from RGustBardon/if-not
Merge pull request #410 from RGustBardon/infinite
cleaner fix for 0/0, 1/0, -1/0 (issue #410)
minor --make stuff
v1.3.1
add the space in “return λ” and similar cases
v1.3.2
revert f5419ec2d6b7e341f99a1f09a63ff6b6a0e3fc95 (squashable directives)
emacs vars
added separate step to discard redundant directives
fix for redundant directives in same scope
"sys" → "util"
be wary of some conditional operators in lift_variables (fixes #435)
fix --lift-vars for "for(var i; ...)" (var declared in for but uninitialized)
Merge pull request #432 from rvanvelzen/fix-431
v1.3.3
I'm officially switching to 4-space indentation.
add note about UglifyJS 2.0
Merge pull request #458 from rvanvelzen/div-with-regexp
discard all \uFEFF characters (close #462)
disallow newlines in strings (fix #463)
Merge pull request #343 from spadgos/patch-1
Reverting "discard all \uFEFF characters (close #462)"
treat \uFEFF as whitespace
v1.3.4
add note about the NPM package, which switched to v2
drop the pledgie link
v1.3.5
Nick Fisher (2):
Added an example of how to use the `defines` option (see #281)
Fixed syntax
Paul Baumgart (16):
Constant fold modulo operator.
Parenthesize sequences in constructor calls.
Parenthesize object literals with no properties.
Parenthesize function expressions.
Don't insert semicolons before multi-line strings.
add semicolon after debugger statements if necessary (closes #353)
add tests for issue 353
fix expected output for issue14 test to have literal tab instead of escaped tab
make note of newlines in comments for ASI purposes (closes #349)
add nodeunit devDependency and add 'npm test' script to package.json
require local version of uglify instead of global
fix consolidator to use new 'directive' AST node
move consolidator tests to test/ dir and run them via 'npm test'
add npm-debug.log to .gitignore
Add base54 digits override for unit tests.
Also descend arg AST nodes in ast_squeeze function-call handler
Quildreen Motta (1):
Adds missing whitespace characters to WHITESPACE_CHARS
Richard van Velzen (4):
Fix for issue #386
Optimize list of digits for generating identifiers for gzip compression.
Fix issue #431
Fix division with regexp as right generating invalid code.
Robert Gust-Bardon (9):
Consolidate null, Boolean, and String values
Fix issues #162 and #271
Implement the idea from issue #325
Fix issues #222 and #324
Revoke #368 and fix issue #383 to make jQuery tests pass
Implement #368 (too many parentheses)
Remove unnecessary "+" sign in SignedInteger
Remove NOT in IfStatement with EmptyStatement (-s)
Fix representation of infinite values
Thomas Mayfield (2):
add package.json for installation via npm
don't die on byte order mark at the beginning of files
Tim Meadowcroft (1):
Added option to define symbols for replacement (cf #define) and improved constant folding for shortcut operators ('&&', '||' and ternary conditional operator)
Ville Lautanala (3):
Encode null characters as '\0'
Fix tests broken after 3cd9829a
Recognize various Unicode whitespace
Whyme.Lyu (1):
Better parser error message for node
Wolfgang Kluge (2):
Keep end position of each token
Test the local instead of the installed version.
cadorn (2):
More ignore rules
Reserved variable names option
chrisk (1):
Support unicode newlines in make_string (make uglifyjs work for the Dragonfly project).
ikokostya (8):
Fixed parse error position: start line number and column number from 1.
Unsafe Object constructor transformations:
new RegExp() => RegExp() (with --unsafe)
new Error() => Error() (with --unsafe)
Unsafe: new Function() => Function()
Fixed #286: Only one variable declaration allowed in for..in loop
Check abort else statement in make_real_if() when t[0] is "stat"
Merge branch 'master' into abort-else
johnjbarton (1):
parameter start arrives 'undefined' or 'null', any falsy value should be zero
satyr (2):
reimplemented #28 in a sane way
unstrict equals operator when operands share the same type
slavic (1):
make 2 symbol variable names range bigger (from 2862 to 3402)
tav (1):
Implemented --define-from-module constant substitution.
thorn0 (1):
Compatibility with the JScript engine, which doesn't support the \v escape sequence
-----------------------------------------------------------------------
No new revisions were added by this update.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-javascript/uglifyjs.git
More information about the Pkg-javascript-commits
mailing list