platform/upstream/rpm.git
11 years agoInclude %{VCS} in rpm -q --info output
Simon McVittie [Wed, 6 Feb 2013 15:04:23 +0000 (15:04 +0000)]
Include %{VCS} in rpm -q --info output

When packages are built from a VCS with an automatically-assigned
release number, checking the commit ID of an installed package is
relatively common and important.

11 years agoMoving some plugin info prints to debug
Elena Reshetova [Mon, 4 Feb 2013 08:14:23 +0000 (10:14 +0200)]
Moving some plugin info prints to debug

11 years agoUpdate to 4.11.0.1 accepted/trunk/20130203.005124 submit/trunk/20130203.005409
Anas Nashif [Sun, 3 Feb 2013 00:54:07 +0000 (16:54 -0800)]
Update to 4.11.0.1

11 years agoUpdate to 4.11.0.1
Anas Nashif [Sun, 3 Feb 2013 00:45:34 +0000 (16:45 -0800)]
Update to 4.11.0.1

11 years agoupdate BR for python package
Anas Nashif [Tue, 29 Jan 2013 13:53:31 +0000 (05:53 -0800)]
update BR for python package

11 years agoUpdate changelog
Markus Lehtonen [Tue, 29 Jan 2013 13:44:25 +0000 (15:44 +0200)]
Update changelog

Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
11 years agoTemporally making loading of a configured plugin optional
Elena Reshetova [Mon, 28 Jan 2013 07:38:08 +0000 (09:38 +0200)]
Temporally making loading of a configured plugin optional

11 years agofixing error in cleanup if magic db can't be loaded [Elena Reshetova]
Anas Nashif [Thu, 17 Jan 2013 14:31:04 +0000 (06:31 -0800)]
fixing error in cleanup if magic db can't be loaded [Elena Reshetova]

11 years agofixing error in cleanup if magic db can't be loaded
Elena Reshetova [Thu, 17 Jan 2013 14:20:02 +0000 (16:20 +0200)]
fixing error in cleanup if magic db can't be loaded

11 years agoEnable security plugin/Smack
Anas Nashif [Thu, 17 Jan 2013 00:25:08 +0000 (16:25 -0800)]
Enable security plugin/Smack

11 years agoAdding security msm plugin
Elena Reshetova [Fri, 4 Jan 2013 10:46:42 +0000 (12:46 +0200)]
Adding security msm plugin

11 years agoMaking pre/post tsm/psm hooks more consistent.
Elena Reshetova [Fri, 30 Nov 2012 16:35:44 +0000 (18:35 +0200)]
Making pre/post tsm/psm hooks more consistent.

-All post hooks take an additional rc argument that indicates general return code from operation.
-All post hooks are always called if correspoding pre hook has been called.
-The return value from post hooks is curently ignored

Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
11 years agoImproving scriptlet related rpm plugin hooks.
Elena Reshetova [Thu, 22 Nov 2012 18:49:27 +0000 (20:49 +0200)]
Improving scriptlet related rpm plugin hooks.

Three new hooks added:
PLUGINHOOK_SCRIPTLET_PRE_FUNC
Called before scriptlet execution
PLUGINHOOK_SCRIPTLET_FORK_POST_FUNC
Called after fork() but before scriptlet execution
PLUGINHOOK_SCRIPTLET_POST_FUNC
Called after scriptlet execution

Currently pre and post hooks are called for externals and internal lua scripts.
post hook is called even if scriptlet execution has failed and
the return code is given as an argument.

fork_post hook is only called for external scriptlets,
because for internal scriptlets no fork() is currently performed.

Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
11 years agoExtending rpm plugin interface, part 1
Elena Reshetova [Wed, 17 Oct 2012 11:55:39 +0000 (14:55 +0300)]
Extending rpm plugin interface, part 1

This change adds a new type of the rpm plugin, called transaction plugin
and a set of initial hooks for this plugin. The hooks are:

PLUGINHOOK_TSM_PRE
    Pre-transaction hook that is called before an rpm transaction begins
PLUGINHOOK_TSM_POST
    Post-transaction hook that is called after an rpm transaction ends
PLUGINHOOK_PSM_PRE
    Pre-transaction-element hook that is called before an rpm
    transaction-element is processed
PLUGINHOOK_PSM_POST
    Post-transaction-element hook that is called after an rpm
    transaction-element is processed
PLUGINHOOK_SCRIPT_SETUP
    Per-script hook that is called once for each rpm mainainers script
    that is present in the package

Each hook is called for every plugin that have this hook registered.
The avaliable transaction plugins can be specified in macros.in via
transaction_plugins element.

Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
11 years agoAdding manifest support for rpmbuild
Elena Reshetova [Fri, 4 Jan 2013 11:46:32 +0000 (13:46 +0200)]
Adding manifest support for rpmbuild

11 years agoMove build related files to rpm-build
Anas Nashif [Mon, 14 Jan 2013 16:54:40 +0000 (08:54 -0800)]
Move build related files to rpm-build

11 years agoreorg files in sub-packages
Anas Nashif [Mon, 14 Jan 2013 16:12:39 +0000 (08:12 -0800)]
reorg files in sub-packages

11 years agobump python-rpm to 4.10.91
Anas Nashif [Fri, 4 Jan 2013 15:10:17 +0000 (07:10 -0800)]
bump python-rpm to 4.10.91

11 years agospec cleanup
Anas Nashif [Fri, 4 Jan 2013 13:53:10 +0000 (05:53 -0800)]
spec cleanup

11 years agocleanup macros
Anas Nashif [Sat, 15 Dec 2012 19:30:24 +0000 (11:30 -0800)]
cleanup macros

11 years agocleanup macros
Anas Nashif [Sat, 15 Dec 2012 19:30:06 +0000 (11:30 -0800)]
cleanup macros

11 years agoupdate to 4.11.0 beta1
Anas Nashif [Sat, 15 Dec 2012 16:55:56 +0000 (08:55 -0800)]
update to 4.11.0 beta1

11 years agofixed lang_package
Anas Nashif [Fri, 14 Dec 2012 16:57:50 +0000 (08:57 -0800)]
fixed lang_package

11 years agopackaging: change some builddeps to be "pkgconfig" type
Markus Lehtonen [Fri, 14 Dec 2012 01:57:32 +0000 (03:57 +0200)]
packaging: change some builddeps to be "pkgconfig" type

For better portability.

Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
11 years agolang_package now accepts arguments
Anas Nashif [Sun, 2 Dec 2012 18:48:41 +0000 (10:48 -0800)]
lang_package now accepts arguments

-n to specify package name
-f to specify lang file name (without .lang)

11 years agofixed package description
Anas Nashif [Tue, 27 Nov 2012 05:15:28 +0000 (21:15 -0800)]
fixed package description

11 years agoadd python-rpm spec
Anas Nashif [Wed, 21 Nov 2012 23:37:03 +0000 (15:37 -0800)]
add python-rpm spec

11 years agoremove file sepolicy from list
Anas Nashif [Wed, 21 Nov 2012 22:01:26 +0000 (14:01 -0800)]
remove file sepolicy from list

11 years agoinclude errno.h
Anas Nashif [Wed, 21 Nov 2012 21:58:38 +0000 (13:58 -0800)]
include errno.h

11 years agoupdate to 4.11 alpha (4.10.90)
Anas Nashif [Wed, 21 Nov 2012 21:45:23 +0000 (13:45 -0800)]
update to 4.11 alpha (4.10.90)

11 years agolibfile-devel -> file-devel
Anas Nashif [Thu, 8 Nov 2012 20:05:53 +0000 (12:05 -0800)]
libfile-devel -> file-devel

11 years agolibfile -< libmagic
Anas Nashif [Thu, 8 Nov 2012 20:04:38 +0000 (12:04 -0800)]
libfile -< libmagic

11 years agoAssume executable libs
Anas Nashif [Sun, 14 Oct 2012 04:09:08 +0000 (21:09 -0700)]
Assume executable libs

11 years agoremove patch from suse
Anas Nashif [Sat, 13 Oct 2012 04:12:29 +0000 (21:12 -0700)]
remove patch from suse

11 years agoRemove some old craft
Anas Nashif [Fri, 12 Oct 2012 15:49:44 +0000 (08:49 -0700)]
Remove some old craft

11 years agoRemove some old craft
Anas Nashif [Fri, 12 Oct 2012 15:45:43 +0000 (08:45 -0700)]
Remove some old craft

11 years agoRemove some old craft
Anas Nashif [Fri, 12 Oct 2012 15:35:00 +0000 (08:35 -0700)]
Remove some old craft

11 years agoadd packaging files
Anas Nashif [Thu, 11 Oct 2012 22:56:20 +0000 (15:56 -0700)]
add packaging files

11 years agodisable dep tracking
Anas Nashif [Thu, 11 Oct 2012 22:53:59 +0000 (15:53 -0700)]
disable dep tracking

11 years agoFind supplements
Anas Nashif [Thu, 11 Oct 2012 22:52:52 +0000 (15:52 -0700)]
Find supplements

11 years agoFile attributes
Anas Nashif [Thu, 11 Oct 2012 22:52:02 +0000 (15:52 -0700)]
File attributes

11 years agoremove languages we do not yet support
Anas Nashif [Thu, 11 Oct 2012 22:49:02 +0000 (15:49 -0700)]
remove languages we do not yet support

11 years agoGStreamer provides, codecs
Anas Nashif [Thu, 11 Oct 2012 22:48:26 +0000 (15:48 -0700)]
GStreamer provides, codecs

11 years agoFont dependencies
Anas Nashif [Thu, 11 Oct 2012 22:47:45 +0000 (15:47 -0700)]
Font dependencies

11 years agoPython dependencies
Anas Nashif [Thu, 11 Oct 2012 22:47:01 +0000 (15:47 -0700)]
Python dependencies

11 years agoDon't complain about a bad md5 sum for repackaged rpms.
Anas Nashif [Thu, 11 Oct 2012 22:45:41 +0000 (15:45 -0700)]
Don't complain about a bad md5 sum for repackaged rpms.

11 years agoPrefer sys/vfs.h, as statvfs stats all filesystems again
Anas Nashif [Thu, 11 Oct 2012 22:45:00 +0000 (15:45 -0700)]
Prefer sys/vfs.h, as statvfs stats all filesystems again

11 years agodebug link
Anas Nashif [Thu, 11 Oct 2012 22:43:32 +0000 (15:43 -0700)]
debug link

11 years agoCreate a debuginfo package for each subpackage.
Anas Nashif [Thu, 11 Oct 2012 22:42:33 +0000 (15:42 -0700)]
Create a debuginfo package for each subpackage.

Originally from opensuse rpm

11 years agoLet debuginfo packages provide the build-id
Anas Nashif [Thu, 11 Oct 2012 22:32:35 +0000 (15:32 -0700)]
Let debuginfo packages provide the build-id

This patch lets debuginfo packages provide build-id like follows:

 debuginfo(build-id) = c63cb23876c5fa85f36beaff58f8557e1bf22517

Users can therefore ask zypper to install the correct debuginfo package
with:

 zypper install -C "debuginfo(build-id) = c63cb23876c5fa85f36beaff58f8557e1bf22517"

11 years agoFirmware dependencies
Anas Nashif [Thu, 11 Oct 2012 22:30:04 +0000 (15:30 -0700)]
Firmware dependencies

11 years agoDo the symbolic links right in the first place
Anas Nashif [Thu, 11 Oct 2012 22:28:50 +0000 (15:28 -0700)]
Do the symbolic links right in the first place

Since brp-symlink relinks symbolic links to enforce a certain policy we
should do it right in the first place. So this patch changes find-debuginfo.sh
scripts behavior to reflect that policy.

Signed-off-by: Jan Blunck <jblunck@suse.de>
11 years agodebugedit canon fix
Anas Nashif [Thu, 11 Oct 2012 22:27:31 +0000 (15:27 -0700)]
debugedit canon fix

11 years agoremove references to removed brp scripts
Anas Nashif [Thu, 11 Oct 2012 22:26:14 +0000 (15:26 -0700)]
remove references to removed brp scripts

11 years agoSplit sources for debugging into separate -debugsource package
Anas Nashif [Thu, 11 Oct 2012 22:24:15 +0000 (15:24 -0700)]
Split sources for debugging into separate -debugsource package

At the moment the -debuginfo package also include the sources where used
to build the binary. The patches moves them into a separate package -debugsource.

11 years agoshort changelog
Anas Nashif [Thu, 11 Oct 2012 22:23:31 +0000 (15:23 -0700)]
short changelog

11 years agoSupport a "first version wins" semantics.
Anas Nashif [Thu, 11 Oct 2012 22:22:40 +0000 (15:22 -0700)]
Support a "first version wins" semantics.

11 years agoProvide rpmqpack
Anas Nashif [Thu, 11 Oct 2012 22:19:09 +0000 (15:19 -0700)]
Provide rpmqpack

a fast way to list all installed packages are
check if some package is installed. This is a hack.

11 years agostrip off the tag
Anas Nashif [Thu, 11 Oct 2012 22:17:10 +0000 (15:17 -0700)]
strip off the tag

11 years agoadding find-supplements scripts
Anas Nashif [Thu, 11 Oct 2012 22:15:50 +0000 (15:15 -0700)]
adding find-supplements scripts

11 years agorpmrc changes
Anas Nashif [Thu, 11 Oct 2012 22:15:28 +0000 (15:15 -0700)]
rpmrc changes

11 years agoAdd distribution
Anas Nashif [Thu, 11 Oct 2012 22:11:03 +0000 (15:11 -0700)]
Add distribution

11 years agoplatform changes
Anas Nashif [Thu, 11 Oct 2012 22:09:49 +0000 (15:09 -0700)]
platform changes

11 years agoModule aliases
Anas Nashif [Thu, 11 Oct 2012 22:06:22 +0000 (15:06 -0700)]
Module aliases

11 years agoMacro adjustments
Anas Nashif [Thu, 11 Oct 2012 22:05:13 +0000 (15:05 -0700)]
Macro adjustments

11 years agoFind debuginfo
Anas Nashif [Thu, 11 Oct 2012 21:56:57 +0000 (14:56 -0700)]
Find debuginfo

11 years agoExclude /usr/share/info/dir from check-files.
Anas Nashif [Thu, 11 Oct 2012 21:47:26 +0000 (14:47 -0700)]
Exclude /usr/share/info/dir from check-files.

11 years agobrp for tizen
Anas Nashif [Thu, 11 Oct 2012 21:43:08 +0000 (14:43 -0700)]
brp for tizen

11 years agoauto dependencies
Anas Nashif [Thu, 11 Oct 2012 21:13:29 +0000 (14:13 -0700)]
auto dependencies

11 years agoAdd support for weak dependencies:
Anas Nashif [Thu, 11 Oct 2012 19:40:31 +0000 (12:40 -0700)]
Add support for weak dependencies:

A) use RPMTAG_SUGGESTS and RPMTAG_ENHANCES to store them.

   This is different to upstream, which uses RPMSENSE_MISSINGOK
   and RPMTAG_REQUIRES/RPMTAG_PROVIDES instead. I chose different
   tags because I wanted to be compatible. The point is that
   applications that don't know about the new MISSINGOK semantics
   will mis-interpret the provides/requires otherwise, which
   I deemed to risky.

B) use RPMSENSE_STRONG to support a "strong" version, "Recommends"
   instead of "Suggests" and "Supplements" instead of "Enhances".

11 years agoSuspend exclusive database lock when scriptlets get called
Anas Nashif [Thu, 11 Oct 2012 19:38:04 +0000 (12:38 -0700)]
Suspend exclusive database lock when scriptlets get called

Suspend exclusive database lock when scriptlets get called, allowing
read access in scriptlets. Only needed for DB_PRIVATE (aka global)
locking.

11 years agoFix global (DB_PRIVATE) lock code
Anas Nashif [Thu, 11 Oct 2012 19:37:11 +0000 (12:37 -0700)]
Fix global (DB_PRIVATE) lock code

Fix global (DB_PRIVATE) lock code: fix recursion counter, retry
failed lock operations for up to 3 minutes.

11 years agoAllow updates to different architectures
Anas Nashif [Thu, 11 Oct 2012 19:35:52 +0000 (12:35 -0700)]
Allow updates to different architectures

Also test architecture in "refresh" test when not colored. This allows
updates to different architectures.

11 years agoExtended condition support
Anas Nashif [Thu, 11 Oct 2012 19:34:48 +0000 (12:34 -0700)]
Extended condition support

This patch supports an extension in the condition evaluation.
If the condition is a format and returns an nonempty string, it is
assumed to be true.
This mechanism is used by the weakdeps patch to filter the
"RPMSENSE_STRONG" flag.

11 years agodb ro interruptable
Anas Nashif [Thu, 11 Oct 2012 19:33:26 +0000 (12:33 -0700)]
db ro interruptable

11 years agoSupport a database-local fsync setting.
Anas Nashif [Thu, 11 Oct 2012 19:32:19 +0000 (12:32 -0700)]
Support a database-local fsync setting.

Needs berkeley db patch.

11 years agocompate name and version
Anas Nashif [Thu, 11 Oct 2012 19:31:15 +0000 (12:31 -0700)]
compate name and version

Also compare the name and not only the version when checking if
two packages are the same. rh#104066

11 years agoConvert output to the current locale.
Anas Nashif [Thu, 11 Oct 2012 19:26:20 +0000 (12:26 -0700)]
Convert output to the current locale.

Assumes utf8 input if the decoding works, otherwise iso-8859-1.
(imported from suse rpm)

11 years agoMake debugedit build without dwarf.h
Anas Nashif [Thu, 11 Oct 2012 19:24:11 +0000 (12:24 -0700)]
Make debugedit build without dwarf.h

11 years agoadd patch 0
Anas Nashif [Thu, 11 Oct 2012 19:22:12 +0000 (12:22 -0700)]
add patch 0

11 years agoAdd db4 integration patch
Anas Nashif [Thu, 11 Oct 2012 19:18:10 +0000 (12:18 -0700)]
Add db4 integration patch

11 years agoupdate spec version
Anas Nashif [Thu, 11 Oct 2012 18:56:39 +0000 (11:56 -0700)]
update spec version

11 years agoadd spec file
Anas Nashif [Thu, 11 Oct 2012 18:55:01 +0000 (11:55 -0700)]
add spec file

11 years agoRestore legacy behavior on %ghost %config erasure, mostly upstream/4.11.0.1
Panu Matilainen [Thu, 31 Jan 2013 14:03:30 +0000 (16:03 +0200)]
Restore legacy behavior on %ghost %config erasure, mostly

- Make rpmfiConfigConflict() consider any existing %ghost %config
  as "modified". This causes them to be left alone on erasure to match
  long-standing rpm behavior (whether the behavior is intended or not...)
  Ghost config testcase passes again, talk about flip-flop. Sigh.
- Legacy behavior had an extra twist which we are intentionally NOT
  restoring: if a non-empty %ghost exists at build-time, its digest
  is stored in the header despite the file not being present in
  the payload. With previous versions of rpm, the contents of an
  existing  %ghost %config would be compared to the one in header and
  resulting in different behavior on erasure: if the contents matched
  that of build-time, the file would be removed, otherwise it would
  be preserved. Ghosts remembering their identity from previous life
  goes a little bit too far into the land of paranormal for me...
(cherry picked from commit 233e4573d1d21a36d06874fea6c594428033802e)

11 years agoAdjust ghost config test case to match legacy behavior
Panu Matilainen [Thu, 31 Jan 2013 13:55:22 +0000 (15:55 +0200)]
Adjust ghost config test case to match legacy behavior

- Rpm has traditionally (at least for the last decade) preserved
  existing %ghost %config file on erasure. Whether this was intentional
  or not, changing the behavior now risks data loss in case people
  are relying on this long-standing behavior.
- This fails now: The behavior changed as (an unexpected) side-effect
  of commit 389b1ab706be8eddba9f00c7084759f670ce96ac.
(cherry picked from commit 4efec66338ef04d2f1ed20cfba84c47145269d0d)

11 years agoRefresh translations again, sigh
Panu Matilainen [Thu, 31 Jan 2013 06:28:54 +0000 (08:28 +0200)]
Refresh translations again, sigh

11 years agoBump version to dot-paperbag aka 4.11.0.1
Panu Matilainen [Wed, 30 Jan 2013 15:23:31 +0000 (17:23 +0200)]
Bump version to dot-paperbag aka 4.11.0.1

11 years agoDont skip shared file/dir create if the other one is a ghost
Panu Matilainen [Wed, 30 Jan 2013 12:07:33 +0000 (14:07 +0200)]
Dont skip shared file/dir create if the other one is a ghost

- Explicitly test for shared file/dir being ghost when deciding whether
  to skip create: we must never skip the creation of a real file/dir
  because of a ghost as the ghosts do not get created.
- This fixes the shared ghost testcase, ie the thing that commit
  a712252392eca75443ca45c10a72873cabec7963 was supposed to fix,
  without side-effects that I can see.
(cherry picked from commit 147a26c5f672ee3421fe057725f6be357031bbb5)

11 years agoRevert the severely broken FA_SKIP on %ghost "fix"
Panu Matilainen [Wed, 30 Jan 2013 12:01:59 +0000 (14:01 +0200)]
Revert the severely broken FA_SKIP on %ghost "fix"

- Reverts the bad half of commit a712252392eca75443ca45c10a72873cabec7963:
  treating files that wont be installed as skipped seems logical enough,
  but rpm disagrees... making %ghosts FA_SKIP on install causes
  upgrades on %ghost %config files to miss the necessary FA_SKIP getting
  set on the erase half via handleInstInstalledFile() and
  rpmfiDecideFate(), causing %ghost %config files to be nuked. Which isn't
  very nice when your /var/lib/rpm/ contents are packaged as
  %ghost %config...
- This fixes the %ghost %config testcase but breaks the shared %ghost
  case in turn.
(cherry picked from commit 471b74bf6e218abb06bb3131a259cdfaec3f08cf)

11 years agoAdd testcase for %ghost shared with a real file
Panu Matilainen [Wed, 30 Jan 2013 11:52:23 +0000 (13:52 +0200)]
Add testcase for %ghost shared with a real file

- Further parametrize conflicttest.spec to allow for different attributes
- Verify that the real file gets created without conflicts, that
  it doesn't get removed when ghost is removed and gets normally removed
  on erase.
- This is the case commit a712252392eca75443ca45c10a72873cabec7963 was
  supposed to fix (which it does but with, ahem, "slight" side-effects)
(cherry picked from commit 0485eab2c96173d1e34e4edb3570e58a693e255e)

11 years agoAdd testcase for %ghost %config behavior
Panu Matilainen [Wed, 30 Jan 2013 11:46:45 +0000 (13:46 +0200)]
Add testcase for %ghost %config behavior

- Further parametrize configtest.spec to allow for different attributes
- Verify pre-existing %ghost %config is preserved on install,
  that it survives upgrades and gets removed on erase.
  This fails now because it IS broken due to commit
  a712252392eca75443ca45c10a72873cabec7963.
(cherry picked from commit 5121ea2151f13e58740b7b4e7dc20cff5f2475fb)

11 years agoPreparing for 4.11.0 final rpm-4.11.0-release
Panu Matilainen [Tue, 29 Jan 2013 06:30:28 +0000 (08:30 +0200)]
Preparing for 4.11.0 final

11 years agoUpdate translations from Transifex
Panu Matilainen [Tue, 29 Jan 2013 06:22:37 +0000 (08:22 +0200)]
Update translations from Transifex

11 years agoDont mess with global user identity in __scm_setup_bzr
Panu Matilainen [Mon, 28 Jan 2013 20:34:11 +0000 (22:34 +0200)]
Dont mess with global user identity in __scm_setup_bzr

- Bzr defaults to global value on "bzr whoami", so we would mess
  up user identity on "%autosetup -S bzr". Oops. Move identity setting
  after repo initialization and set the identity per-branch, not global.
  The notion of "per-branch identity" seems odd, but then bzr is...
  I guess this would be bzr-speak for "per-directory identity"
(cherry picked from commit 58f41cb4e1721d410017732a55613056b6acb24c)

11 years agoSetup git identity in __scm_setup_git (RhBug:905002)
Panu Matilainen [Mon, 28 Jan 2013 20:31:30 +0000 (22:31 +0200)]
Setup git identity in __scm_setup_git (RhBug:905002)

- For some reason on my own account git keeps functioning even if
  I remove its configuration, but on a newly created account git commit
  bails out with "Please tell me who you are" unless we set the
  (per-repo) configuration here.
(cherry picked from commit 321a87d789bb090c88c541f7f158ae922ddee3ff)

11 years agoSplit __scm_author to separate name and email address macros
Panu Matilainen [Mon, 28 Jan 2013 20:28:35 +0000 (22:28 +0200)]
Split __scm_author to separate name and email address macros

- At least git wants name and email separately, change __scm_author
  to combine the two instead. Related to RhBug:905002.
(cherry picked from commit 1a21dc867dfbf48169782c6b8dbd0e0476e2bbbf)

11 years agoAcommodate to glibc 2.17 __secure_getenv() renaming
Panu Matilainen [Mon, 28 Jan 2013 16:49:01 +0000 (18:49 +0200)]
Acommodate to glibc 2.17 __secure_getenv() renaming

- glibc 2.17 renames __secure_getenv() to secure_getenv() so now we
  need to test for both. Meh.
(cherry picked from commit 49519f2564777fdcfb801088fb3025b3d6f5783a)

11 years agoBe more careful about skipping shared file/directory creation
Panu Matilainen [Mon, 28 Jan 2013 10:50:39 +0000 (12:50 +0200)]
Be more careful about skipping shared file/directory creation

- Only skip shared file/dir creation if its actually being created
  by another package. Previously we could've decided to skip an entry
  where the other file was also being skipped for some other reason.
- Ensure %ghost entries always have FA_SKIP on install, previously
  they often were FA_CREATE which makes no sense for %ghost. We dont
  encounter %ghosts at all during install in the FSM as they dont
  exist in the payload, but the file disposition calculations need
  to take them into account now that we're avoiding redundant creates.
- Fixes a regression from commit f7f5f88f9f3d6587e747b034ccb64a3f00ff4e1e
  which unearthed the %ghost FA_CREATE issue, reported here:
  http://lists.fedoraproject.org/pipermail/buildsys/2013-January/004047.html
(cherry picked from commit a712252392eca75443ca45c10a72873cabec7963)

11 years agoAdjust rpmspec -P to invoke non-build parse
Alexey Tourbin [Tue, 22 Jan 2013 00:47:15 +0000 (00:47 +0000)]
Adjust rpmspec -P to invoke non-build parse

In the query mode, rpmspecQuery puts the parser into non-build mode
by invoking rpmSpecParse with (RPMSPEC_ANYARCH|RPMSPEC_FORCE) flags.
The same flags should be applied in the parse mode, so that it is
possible to preprocess specfiles separately, i.e. without also
installing sources and patches.

Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
(cherry picked from commit 4eb824937ce536aeecdfdcab2d88e28083bba85b)

11 years agoMake double-quoting work for special %doc (and %license) too
Panu Matilainen [Tue, 22 Jan 2013 05:55:11 +0000 (07:55 +0200)]
Make double-quoting work for special %doc (and %license) too

- Up to now, special %doc has been different from everything else in
  %files: double-quoting which is used to escape eg spaces in filenames
  has not worked, but single-quoting and backslash-escapes (which do
  not work elsewhere in %files) "worked" by happenstance due to getting
  passed verbatim to shell/cp. Those and various other %doc hacks people
  have come with stopped working (ticket #858) as starting from
  commit 29677605d44dc9cba3119135653ba0372ab58037 we perform the
  copies in slightly more controlled manner.
- Rather than re-enable old quirks, make %doc and %license behavior
  consistent with the rest of the %files section: double-quoting and
  globs work, other escaping methods do not.
- This does mean a minor (as docs with spaces are relatively rare)
  compatibility rift in specs, the "official workaround" is that
  if compatibility with older rpm versions is required globs can
  be used instead of quotation.
(cherry picked from commit a1d9364adb556813886d91b2799217a412ac5bb0)