[Pkg-freeipa-devel] oddjob: Changes to 'refs/tags/debian/0.34.3-1'
Timo Aaltonen
tjaalton at moszumanska.debian.org
Wed Oct 21 19:51:22 UTC 2015
Tag 'debian/0.34.3-1' created by Timo Aaltonen <tjaalton at debian.org> at 2015-10-21 16:12 +0000
tagging package oddjob version debian/0.34.3-1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=LCkU
-----END PGP SIGNATURE-----
Changes since the dawn of time:
David King (1):
Update dbus-send dependency for new dbus (#1170584)
Nalin Dahyabhai (107):
- avoid defining DBUS_MAJOR_VERSION when we don't have to
* src/oddjob-dbus.c: fix a quoting warning in a comment. Don't use
- set the flags in the final method entry, not that it should matter
- fixup remaining namespace thinkos
- update to version 0.29.1 to include config file fixes and a couple of bug
- bump release to 1
- bump release to 0.29.1
- add targets for an 'archive' target
- fix 'archive' target
* src/oddjobd.c(oddjobd_exec_method): provide useful error messages
* src/pam_oddjob_mkhomedir.c,src/pam_oddjob_mkhomedir.8.in: try to
- don't refer to CVS in the docs since we've moved to git
* src/launch.c,src/mainloop.c,src/mkhomedir.c,src/mounthomedir.c,
* scripts/oddjobd.in: advertise the verbs we recognize (#522131)
* Makefile.am: don't try to use RPM for tagging; take the release
* src/Makefile.am, src/oddjobd-mkhomedir.conf.in: move the helper
- also don't instlal the header file
* oddjob.spec,configure.ac: bump version to 0.30
* src/Makefile.am: make the library and headers conditional
- make installation of the library and header conditional
- rpmlint cleanups
- some rpmlint cleanups
- fixup the oddjobd config for mkhomedir to account for the helper moving
- fix a typo in %%post
- try to make the configuration file look the same across multilib arches
- conditionalize the building of PAM-specific bits
- move some files around that we don't actively build any more or yet
- more makefile tweaks
- merge some history from Fedora
* src/oddjob_dbus.c: refuse to try to close the connection to the
* src/oddjobd.c: report ENOENT if that's why getfilecon() fails
* src/sanity.sh,tests/sanity.sh.in,tests/{007,008}/expected_stdout.in:
- another shot at fixing timestamp-based conflicts
- bump the release number
- more to do. always more to do.
- fix a typo in the summary
- assume fedorahosted hosting's going to be set up
- import some Fedora history
- fix that typo in the description while we're in here
- aaagh. macros in changelogs. i just know i'm gonna get dinged for that.
- actually set a value for DBUS_INTROSPECTION_INTERFACE and
- bump version for this fix
* configure.ac: fixup the permissions on tests/sanity.sh so that tests
- note the bug number from Red Hat bugzilla
* src/oddjob_dbus.c: just stop calling dbus_connection_close(), which didn't previously cause us to abort() before, but does now (#634356)
- try to SIGHUP the messagebus daemon at first install so that it'll let us claim our service name if it isn't restarted before we are first started (same as #636876)
when the mkhomedir helper has to create intermediate directories, don't apply a umask that might have been supplied on its command line (#666418)
set permissions on intermediate directories correctly when the mkhomedir helper has to create some, using values from /etc/login.defs when no default umask is specified
juggle the umask correctly when creating intermediates
- merge .spec with raw hide's
- clean up .spec file changelog order
- make the init script exit with status 2 when given an unknown command, rather than with status 1 (#674534)
- also tell the system message bus to reload its configuration when we install a subpackage with a new service in it
- nope, that script change is not packaging, so bump the version to 0.31.1
notice when we fail to register a name with the message bus
init scripts / unit files
add internal logic for sending replies as signals
systemd updates, bump version
clean up systemd logic
conditionalize scriptlets on init system
add --enable-pie and --enable-now configure flags (#852800)
- don't worry about moving things from /usr to / when they're the same (#852800)
correct the release with no-separate-usr
- add missing requires(post) on killall, which we use to poke the message
fix use of incorrectly-named facilities (#802719)
install a systemd unit file on in-devel releases
tag 0.31.3
merge fedora spec file updates
resync with Fedora's .spec file
fix the NAME values to make lexgrog stop barfing
fix the NAME values to make lexgrog stop barfing
fix the unit file, tag as 0.31.4
fix a skewed date
Add some man pages
- require "dbus" at the package level (#1085450)
Allow the oddjob_request timeout to be controlled
Stop explicitly setting a umask for mkhomedir
Fix crashes in pam_oddjob_mkhomedir
Call getpwnam_r() correctly
Add some notes on the helper's options
Remove fixed limits on reply sizes
Note that the PAM module takes no options
Resync .spec file with Fedora
Make the dbus-send move depend on the OS release
Remove an unused variable
Fix warnings about passing size_t to fprintf
Remove an unused variable
Search for services after checking the sender
Improve matching against defined services
Move to using per-service connections
Exit if we printed a help message
Don't leak SELinux range storage on cleanup
Tell the bus daemon to allow introspection
Increase the initial size of the PAM module's buf
Document other attributes of "helper" nodes
tag 0.34
Disable maintainer rules by default
autogen.main: pick up LDFLAGS from rpm-build
Save and restore LDFLAGS while performing tests
Make --enable-now actually work (whoops)
Use -fPIE when linking binaries rather than -fPIC
tag 0.34.1
Add a test for larger reply messages
Correctly initialize the reply pointer (#1236970)
tag 0.34.2
Register service names after hooking up dispatcher
tag 0.34.3
Timo Aaltonen (4):
initial packaging
close ITP
remove obsolete files not in tarballs
releasing package oddjob version 0.34.3-1
nalin (328):
initial import
add archive/tag/force-tag targets
use a "make dist" tarball as source0
- add 'autogen' script
update the pkgbuilddir
fix pam_oddjob_mkhomedir not being useful
0.4, because... why not.
log debug message for each request
call the method correctly
handle empty output buffers correctly
0.5
- use oddjobd-local.conf instead of oddjob-local.conf
pam_oddjob_mkhomedir: forward text messages to the user
fixup garbage derefs
- define mydatadir
- new app: dbus-launch
- screen for characters which wouldn't work being fed to helpers line-at-a-time
namespace syslogs
- use DBUS_MAGIC1 where it makes sense
and look in /etc/oddjobd.conf.d for .conf files
- fix deny/allow processing to match docs
- beginnings of "mount my homedir for me" stuff
- fix use of man macros
- placeholders for mounthomedir stuff
- placeholder
- use $localstatedir/lock for holding refcounts and locks
- fine-tune the method by which we derive lock names
- add userfstab to the list of installed files
include userfstab in distball
include /etc/PACKAGE in binary RPMs
- fix comments
- don't leak the config file and lock descriptors
- set the fscreatecontext when creating directories, too
- don't leak the args list
- syslog errors
- remove redundant setting of creation contexts around calls to my_mkdir()
- note that both mkhomedir and mounthomedir use selinux.h
- include the method name when we note that a helper exited with an error code
- call matchpathcon_init
- exit with failure at the first error
- don't return an error code on success
- ignore SIGPIPE instead of defaulting for it
- hint that the relative destination directory is optional, which it is
- create a shared oddjob_selinux_mkdir() function, using it instead of
- clean up some warnings
- don't bail on the first comment, duh!
- expect selinux, but check if matchpathcon_init() exists
- add new helper
- rework configuration file format to allow specification of multiple objects
- make "list" more verbose and multi-object aware
fix "dist" target
initial commit
rename MAGIC_DBUS1 to DBUS_PACKAGE
*** empty log message ***
a bit of comment on the config files
- switch to proper copyright notices
add python bindings for the client library
short notes on deployment
- beginnings of a generic lookup-stuff service backend
add proper dbus requirement
- include more files in the default tarball
- add a means of determining the default object and interface names
- add more things i need to do before i call this 1.0
- make oddjob_dbus_send_message_response_success take a variable number of
more additions
more updates
- echo out the command-line arguments we got
- add the "argument_passing_method" attribute for helpers, and heed it
- change signatures of oddjob_dbus_call_method/oddjob_dbus_call_methodv to
- catch-up to the recent change in oddjobd
- account for changes in the caller APIs
- account for changes to APIs
- exercise "argument_passing_method" by splitting sanity up
- new milestone
- clear output buffers in case we don't get anything
- wait() on exec failures so that we don't leave zombies
- add a really brief example
- add the sample python script as a doc file
- oops, need this
- add sample helpers, create the included config directory
- more items
- add printenv.sh to exercise upcoming call information in environment code
- add more scripts for testing
- define names for environment variables
- stop trying to make output prettier
- add the ability to make object paths, interface names, and method names
- document wildcard object paths
bump version
remove more debug spew
- use a preprocessor define to set the flags we use for matching user names
- start abstracting things so that we can add our own watch types
- add oddjob_strdup_printf()
- move test-only helpers to a sample local file
- add oddjob-specific watches
- generic buffer-of-bytes management
- small cleanups
- handle helper requests asynchronously
- rework library module setups so that linking continues to work
- don't add compile flags to SELinux-specific CFLAGS, because they'll wind up
- build without SELinux first, then with, only using the latter build, so
- make note of how the system message bus's APIs can be called
- track the sender's SELinux context in oddjob_dbus_message structures
- use the public header. this is a sample, after all.
- track whether or not SELinux is enabled/enforcing
- make some code #ifdef SELINUX, because it refers to data which already was
- preprocess the DTD depending on whether or not we support SELinux
- nothing SELinux-specific about the function signature
- notes on selinux additions
- don't strlen() a possibly-binary string
- fix it so that SELinux-specific bits aren't in the man pages unless we
bump version
force a shorter timeout if we have no watches but are checking PIDs
Avoid calling a function which didn't exist yet in 0.33.
bump version
make internally-defined methods also return an empty error text field on success
remove the SELinux todo item, it is done
- actually try to launch the service as a daemon
- fix typo, patch from Mark McLoughlin
- don't install the mounthomedir stuff -- it's not fully cooked yet
- add sample data for a sample service
- bump the version number, handle sample data
stop installing things we don't use
- support multiple service names
fixup some warnings
- update to support multiple services
- spit up on configuration file invalidity
- add longer-form docs, get rid of StudlyCaps
- oops, committed the wrong file
- remove wildcarding of interface and method names, which is sorta bogus
build long-form docs for packages
bump version number to 0.14
- try to split up whether or not we support SELinux in our ACLs from whether
- use xmlto to generate HTML, and include it in tarballs
update python binding for new any-service state of things
- add an XML comment in the oddjobd config sample
force people to request reformatting of the docs explicitly
- a maintainer-options autogen
try another way to include HTML docs in dist tarballs
stop forcibly regenerating the docs
add auto-generated html
- add support for multiple busses
- add a test framework, using a session bus to run tests
add testing framework
commit the test driver
create the pidfile before changing directories, in case it's a relative path
- don't bother clearing the supplemental list unless we're on the system bus
- add more options to the test set to handle different versions of D-BUS
remove newline which should not be there
fix extra spaces problem
- use the new names for well-defined information we need for connecting to the
- hey, who loves auto-generated files?
- match error name style of the regular bus
- update to match current code changes
- open the file before we daemon(), fdopen it after
localize libxml a bit more
- update tests
- start down the road to being able to specify ACLs anywhere
- ACLs for everybody!
- stress more strongly that we don't authenticate users
- source code cleanups
- raise that completely-arbitrary limit
- parse command-line arguments in helper definitions
- update expected output to match new error text names
update auto-generated text
fix transposition
update auto-generated text
- use argument parsing in the service launcher
- minor fixes here and there
- more autogenerated stuff
fix ws
- add production of introspection data (at least for services which don't
* src/oddjob_dbus.c: DBUS_NAME_FLAG_PROHIBIT_REPLACEMENT is the default, so
* oddjob.spec: modify to conform more closely to the Fedora Packaging Guidelines
- pre-tag commit
- include the readme file in distballs
* src/mainloop.c,src/mainloop.h(mainloop_reinit): add, doing what
* src/oddjob_dbus.c: track the bus type and reconnect timeout we were given at
* src/oddjobd.c(main): call reconnect_if_needed before every mainloop iteration.
* scripts/oddjobd.in: set the reconnect timeout to 5 minutes.
* src/pam_oddjob_mounthomedir.c: add a "don't use this" comment. Seriously,
* configure.ac: add hard-coded definitions for reconnect-related timeouts.
* src/oddjobd.8.in: add notes about the -t option and restarting.
- scratch off one item
reformat, commit
tag 0.18-1
reformat, commit
* src/oddjob_dbus.c(oddjob_dbus_send_introspection_text): add, for
* src/oddjobd.c: add a pile of introspection-related format and constant
* configure.ac: add hard-coded definitions for introspection-related names.
- mark off TODO item "introspection"
reformat, commit
2005-12-21 nalin
* src/oddjobd.c: if we don't provide methods, pretend there's no such object,
- update
* src/pam_oddjob_mkhomedir.8.in: mention the interface name.
- update
- tag to 0.20
add a URL, duh
* autoversion.in: add, so that configure and the .spec file hopefully don't
* Makefile.am: compare autoversion with the .spec file version.
reformat, update
tag 0.20-1
- update
reformat, update
- update year
* src/oddjob_dbus.c (message_has_path): abstract out, for the sake of versions
* src/oddjob_dbus.c (oddjob_dbus_connection_close): add an abstraction to avoid
- explicitly check for a C compiler
* src/oddjobd.c (main): don't try to set a reconnect timeout until after we're
* src/oddjob_request.c (main): add a -I flag to perform introspection.
* src/oddjobd.c (main): don't try to set a reconnect timeout until after we're
- pass command-line arguments directly to configure
* sample/oddjobd-sample.conf.in: fix pkglibexecdir confusion (#176452).
reformat
- 0.22
- bump version as known to autotools
- reformat
* src/oddjob_dbus.c(oddjob_dbus_get_selinux_context): handle
* doc/oddjob.xml.in: mention known implementation limits.
* configure.ac: when checking if the message bus supports the
tag 0.23-1
oops, need to changelog that
rebuildx2
* src/oddjob_dbus.c(message_has_path): dbus_message_has_path was
reformat
prepare to increment version, require xmlto
force --enable-xml-docs on, which is why we need to require xmlto
tag 0.24-1
* doc/oddjob.xml.in: make use of DocBook 4.4 vs. 4.3 a configurable
* src/oddjobd.c(oddjobd_introspect_method): apply ACLs to the
* src/util.c(oddjob_resize_array): correctly handle cases where the
* configure.ac: make the names of introspection-related interfaces
* src/oddjob_dbus.c(oddjob_dbus_unbind): wrap the release_name()
* src/oddjobd.conf.in: allow root to call "reload" by default.
* src/oddjobd.c: make the name of the configuration file a global,
* src/mainloop.c(mainloop_reset_signal_handlers): don't reset SIGHUP.
* src/mainloop.c(mainloop_reset_signal_handlers): don't reset SIGHUP.
* src/mkhomedir.c: add -u and -s flags to override the umask and
* src/oddjobd.8.in: note that SIGHUP works now.
* src/oddjobd.conf.5.in: add the missing XML document type identifier
* QUICKSTART: remove notes that oddjobd doesn't do restarts well.
* TODO: remove graceful reload from the list
* use quotes around interface names, which the default stylesheet doesn't
- updates
* configure.ac: print the version of D-BUS which we're using
* scripts/oddjobd.in: run "oddjob_request reload" to reload instead of
* src/make-i11nconf.sh.in: add a script to generate all of the needed
* src/oddjobd.conf.in: remove introspection ACL entries.
* src/Makefile.am: use make-i11nconf.sh to generate a useful
* tests/Makefile.am: fix 'distcheck'
* src/oddjobd.conf.5.in: note that oddjobd provides introspection if
* src/oddjobd.c(oddjobd_introspect_method): provide a list of objects
* oddjobd.spec: own /etc/oddjobd.conf.d/oddjobd-introspection.conf
* scripts/oddjobd.in: switch to the more conventional SIGHUP to force
* oddjob_dbus.c(oddjob_dbus_listener_remove_method): once we find the
* configure.ac,src/oddjob_dbus.c: move the warning about a missing
* src/oddjobd.c(oddjobd_introspect_method): report the right
- tag 0.25
- add recommended dependency on pkgconfig in the -devel subpackage
* make sure that eightx.sh and tenx.sh are included in the dist tarballs
* if the user requested introspection data, then always return 0 instead of
* add macros for "unimplemented method" error, and a variable name for the
* src/oddjob_dbus.c(oddjob_dbus_filter): check for cases where we
* src/oddjobd.c(object_has_method): don't get fooled by methods which
- add more TODO items
* tests/*/description: add descriptive names for the various tests
- ensure that the tarball includes the newest tests
- more TODO items, this time for the test cases
- note $ODDJOB_CALLING_USER
- 0.26
- the output now includes the user name, so we have to generate it
add tests for introspection
* src/oddjob_dbus.c: don't explicitly subscribe to any messages. This
- make the close/disconnect wrapper globally-callable, because D-BUS^WD-Bus
* src/oddjob_request.1.in,src/oddjobd.8.in,src/oddjobd.c,
- update docs and comments because D-BUS is now called D-Bus
rebuild
- rebuild
- rebuild
- rebuild
- rebuild
- unmark the init script as a %config file (part of #197182)
* configure.ac: fix saving/restoring of CFLAGS while we're trying to
* oddjob.spec: disable dependency tracking at configure-time (#228928)
* src/mkhomedir.c: use strtol() instead of atoi() to parse the umask,
* src/oddjob_request.c: handle the more-natural method of specifying
regenerate
* src/launch.c: correct a compile warning about use of parentheses
* configure.ac,src/common.h,src/mainloop.c,src/oddjob_dbus.c: pick out
* src/pam_oddjob_mkhomedir.8.in: note that the umask and location of
- 0.28-1
* src/oddjob_request.c: sanity-check that we have a method name
* src/mkhomedirfor.in: wrapper for dbus-send which calls mkhomedirfor
* src/mkmyhomedir: wrapper for dbus-send which calls mkmyhomedir
* src/reload: wrapper for dbus-send which calls our reload method
* src/oddjob{,d}.conf.in,src/oddjob{,d}-mkhomedir.conf.in: split the
* configure.ac,src/common.h: hard-code the introspection interface
* src/oddjobd.c: implement a "listall" which bypasses ACLs so that
* src/pam_oddjob_mkhomedir.{c,8.in}: account for the new service naming.
* autogen,autogen.maint: try to guess lib/lib64 correctly
* src/oddjob_dbus.c: never call dbus_message_iter_get_array_len().
* src/oddjobd.conf.5.in: fix formatting so that making the SELinux-
- rename to note that we don't use the struct FTW
* src/mainloop.c(timeout_dbus_toggle,handle): prefer to call
- adjust the comment we add to the top of the generated file
* doc/oddjob.xml.in: note what kinds of data to expect back from
* oddjobd.c: include -t in the terse usage message
- split out mkhomedir bits, add wrapper scripts for dbus-send
* scripts/oddjobd.in: work on conforming to Fedora initscripts
* src/oddjob_dbus.c: don't bother attempting to determine the SELinux
- hard-code the introspection interface and method names, because configure.in
* scripts/oddjobd.in: output a message when we 'reload'.
- give 'status' a full path to chew on
- fixup for current ls
- make sure sanity.sh gets built
- strip out the block count at the start of 'ls' output
- strip that spurious newline
- remove spurious newline
- updates
- bump the package version as well
- regenerate
- bah, use the dist tag
- fix tagging in light of a dist tag
- remove -mkhomedir->pkgconfig dependency, that's just silly
- make sure that the octal help is always preceded by a "0"
- oddjob_mkhomedir: apply the umask to files as well as directories (#246681)
- fix the interface name for the mkhomedir methods, so that they match the
- fix the comment at the top, too
More information about the Pkg-freeipa-devel
mailing list