[fedora-india] Fwd: [Rpm-announce] rpm 4.6.0 release candidate 4

Rakesh Pandit rakesh.pandit at gmail.com
Fri Jan 30 11:14:35 UTC 2009


---------- Forwarded message ----------
From: Panu Matilainen <pmatilai at redhat.com>
Date: 2009/1/30
Subject: [Rpm-announce] rpm 4.6.0 release candidate 4
To: rpm-maint at lists.rpm.org, rpm-announce at lists.rpm.org

Here comes what's hopefully the last RC before 4.6.0 final, been dragging
on too long already. Download info and generic 4.6.0 release note draft is
(still) at http://rpm.org/wiki/Releases/4.6.0, below you'll find the
change summaries since RC1 - if you wonder what happened to RC2 and RC3,
they went online unannounced, trying to make up for that here...

Changes from RC3 - RC4:
- package having explicit RPMTAG_SOURCEPACKAGE is always SRPM (rhbz#478907)
- safer update on file digest algorithm change (#479869)
- dont think of gpg-pubkeys as source headers
- reflect Filemd5s -> Filedigests index name change in macro config too
- dont segfault on NULL argument to rpmpsGetProblem()
- dont segfault on "" rcfiles argument
- dont segfault if rootDir is NULL
- dont try to create "" directories
- dont segfault in pgpPubkeyFingerprint() if rpmDigest* fails
- fix regression in rpmdsMerge() on rpmdsSingle() and rpmdsThis() items
- fix memleak on I18N strings retrieved with HEADERGET_ALLOC
- use temporary table for Depends in sqlite backend too
- dont run %pre- and %posttrans scripts if --noscripts is specified
- delay NSS initialization to limit problems with forking (rhbz#476737)
- permit --noscripts on verify
- javadeps format string fix
- fix buffer overlow on long filelist entries (susebz#397006)
- check for dbenv->failchk() exit code
- check for chroot() result in psm too
- check for termination signals during query loops
- added ARMv7 support
- added ISA macros for ARM
- update documentation wrt gpg path macro (rhbz#476201)
- clarify documentation wrt generating API/hacking documentation
- dont bother with translations for internal/test-only tools
- updated Swedish translation
- updated Polish translation

Changes from RC2 - RC3
- make sure headers can be included from C++ (ticket #12)
- work around jrpm (?) produced packages not having RPMTAG_SOURCERPM
- empty arrays on queryformats return nothing at all
- fix segfault on packages without scriptlet program tag (rhbz#475214)
- fix pkg-config provide generation on interdependent .pc's (rhbz#473814)
- tighten up setting and checking of buildroot (ticket #10)
- bunch of tag extension regression fixes from RC2 (rhbz#474550)
- fix python return type for non-existent tags (rhbz#473239)
- backwards compatibility include for rpmGenPath() (rhbz#473420)

Changes from RC1 - RC2
- read default popt aliases again (rhbz#473167)
- fix old, pathological memory fragmentation on fingerprinting (rhbz#472507)
- fixes to permit relative paths on buildtree creation
- restore SIGPIPE to default action for scriptlets (rhbz#471591)
- calculate total file size correctly in build (mdvbz#45820, rhbz#247374)
- tag extension enhancements generalizing numeric data handling (rhbz#471820)
- use better hash to improve initial install speed
- filter out mtime difference of shared files on verify (rhbz#426672)
- fix segfault on relocation (rhbz#468319)
- fix old memleak in rpmtsAddInstallElement()
- make noarch subpackages actually work
- skip adding identical NEVR to transaction on install too (rhbz#467822)
- permit tab as whitespace between macro arguments (rhbz#467567)
- adjust for file/libmagic output change (rhbz#468129)
- SH-platform fixes (ISA-macros, IEEE-compliance)
- PATH_MAX portability fixes (ticket #3)
- avoid leaking C99 options into stock %{__cc} macro
- remove extranuous trailing slash from pkgconfig includedir
- make peace with new libtool
- updated Brazilian Portugese translation
- updated Serbian translation

Still reading? Quite a pile there, I agree. In retrospective, RC3
should've probably gone out as 4.6.0 but it's a bit too late to speculate
on that. Oh and FWIW, a big part of the fixes that went to especially RC4
were found by Pixel through testing on Mandriva rpm-software stack,
exercising different code paths than what typically gets tested by me so
it would've been hard to notice them otherwise, thanks for spotting and

RC4 will go out as 4.6.0 final as-is in about a week, unless severe
regressions from RC3 are found.

       - Panu -

Rpm-announce mailing list
Rpm-announce at lists.rpm.org


More information about the Fedora-india mailing list