Panu Matilainen [Thu, 8 May 2008 10:40:26 +0000 (13:40 +0300)]
Oops, matchpathcon() wants st_mode, not the entire struct stat
Panu Matilainen [Thu, 8 May 2008 10:35:27 +0000 (13:35 +0300)]
Block signals while moving the rpmdb to final location after rebuild
- interrupt there would be pretty disastrous..
Panu Matilainen [Thu, 8 May 2008 10:30:16 +0000 (13:30 +0300)]
Lose the silly db argument from block/unblockSignals()
- unused, and unusable anyway as signals are global, not per db thing...
Panu Matilainen [Thu, 8 May 2008 10:06:34 +0000 (13:06 +0300)]
Restore rpmdb selinux contexts to default after rebuild (rhbz#442149)
Panu Matilainen [Thu, 8 May 2008 09:40:37 +0000 (12:40 +0300)]
is_selinux_enabled() returns 0 if not enabled, not -1
Panu Matilainen [Thu, 8 May 2008 08:53:42 +0000 (11:53 +0300)]
Check for transaction root sanity in rpmtsSetRootDir()
- only regular paths starting with / permitted
- change prototype to permit return codes
Panu Matilainen [Thu, 8 May 2008 08:38:52 +0000 (11:38 +0300)]
Set ts rootdir to /, not empty if rootDir arg is NULL
- lets see what break... ;)
Panu Matilainen [Thu, 8 May 2008 08:27:58 +0000 (11:27 +0300)]
Only append trailing slash to ts rootdir if rootdir != "/"
Panu Matilainen [Thu, 8 May 2008 07:20:39 +0000 (10:20 +0300)]
Fix up some fuzz in Finnish translation
Panu Matilainen [Thu, 8 May 2008 07:07:31 +0000 (10:07 +0300)]
Fix up / remove some fuzzy German translations
Panu Matilainen [Thu, 8 May 2008 07:00:04 +0000 (10:00 +0300)]
Updated Serbian translations from Igor Miletic (rhbz#441101)
Panu Matilainen [Thu, 8 May 2008 06:56:54 +0000 (09:56 +0300)]
Updated Polish translation from Piotr Drąg
Panu Matilainen [Thu, 8 May 2008 06:53:36 +0000 (09:53 +0300)]
Updated German translation from Fabian Affolter (rhbz#438178)
Panu Matilainen [Thu, 8 May 2008 04:48:51 +0000 (07:48 +0300)]
Check rpmAddSignature() return codes (rhbz#442761)
- headerGetEntry() on RPMTAG_HEADERIMMUTABLE already caught the corrupted
package, we just didn't paying attention
- make the hge failed error message a bit more meaningful
Panu Matilainen [Wed, 7 May 2008 09:15:22 +0000 (12:15 +0300)]
Oops, don't access variable after freeing
Panu Matilainen [Wed, 7 May 2008 08:33:33 +0000 (11:33 +0300)]
Oops, plug leak introduced by previous commit
Panu Matilainen [Wed, 7 May 2008 08:12:17 +0000 (11:12 +0300)]
Simplify doPatch()
- use %{uncompress} macro to pass files to patch always
- gains support for lzma patches for free
Panu Matilainen [Wed, 7 May 2008 06:48:39 +0000 (09:48 +0300)]
Plug another memleak from genCpioListAndHeader()
- free fuser + fgroup info instead of NULLing, add comment on side-effects
Panu Matilainen [Wed, 7 May 2008 05:32:36 +0000 (08:32 +0300)]
rpmfi actions is already alloced by rpmfiNew(), don't leak memory
Panu Matilainen [Tue, 6 May 2008 09:48:09 +0000 (12:48 +0300)]
Eliminate nasty hackery with "special" doc + docdir format
- figure out docdir early on in spec parsing, store in package struct
instead of abusing parse buffer
- fixes memleaks from docdir_fmt and headerSprintf too
Panu Matilainen [Tue, 6 May 2008 07:00:07 +0000 (10:00 +0300)]
Eliminate static buffer in lua findkey()
Jindrich Novy [Tue, 6 May 2008 06:47:18 +0000 (08:47 +0200)]
Nuke EXPRBUFSIZ, it is now useless
Jindrich Novy [Tue, 6 May 2008 06:46:50 +0000 (08:46 +0200)]
Don't use static buffer for parsing an expression
Panu Matilainen [Mon, 5 May 2008 10:52:19 +0000 (13:52 +0300)]
Indentation fixes
Panu Matilainen [Mon, 5 May 2008 08:53:48 +0000 (11:53 +0300)]
Eliminate static sized string buffer in parseScript()
Panu Matilainen [Mon, 5 May 2008 08:37:08 +0000 (11:37 +0300)]
Trap RPMRC_FAIL returns from spec parse with abort() for now
Panu Matilainen [Mon, 5 May 2008 08:31:09 +0000 (11:31 +0300)]
Make readLine() return PART_ERROR on errors
Panu Matilainen [Mon, 5 May 2008 08:27:49 +0000 (11:27 +0300)]
Fix (hopefully ;) signaling for line continuations on spec parsing
Panu Matilainen [Mon, 5 May 2008 08:20:13 +0000 (11:20 +0300)]
Indentation fixes
Panu Matilainen [Mon, 5 May 2008 08:15:12 +0000 (11:15 +0300)]
Make parseSpec() return PART_ERROR on all errors
- also fix memleak from buildroot errors
Panu Matilainen [Mon, 5 May 2008 08:00:45 +0000 (11:00 +0300)]
Make parseFiles() return PART_ERROR on failure
- streamline exits by assuming failure
Panu Matilainen [Mon, 5 May 2008 07:56:55 +0000 (10:56 +0300)]
Make parseScript() return PART_ERROR on errors
- streamline exists by assuming failure
Panu Matilainen [Mon, 5 May 2008 07:46:43 +0000 (10:46 +0300)]
Make parseDescription() return PART_ERROR on errors
Panu Matilainen [Mon, 5 May 2008 07:42:41 +0000 (10:42 +0300)]
Make parseChangeLog() return PART_ERROR on errors
- streamline exits, fixing memleak from stringbuf on various paths
Panu Matilainen [Mon, 5 May 2008 07:34:33 +0000 (10:34 +0300)]
Make parseBuildInstallClean() return PART_ERROR on errors, streamline exits
Panu Matilainen [Mon, 5 May 2008 07:24:22 +0000 (10:24 +0300)]
Make parsePrep() return PART_ERROR on errors
Panu Matilainen [Mon, 5 May 2008 07:08:33 +0000 (10:08 +0300)]
Streamline parsePreamble() exists
- assume failure, single point of exit
- fixes memleak from NVR
- use PART_ERROR for error exits
Panu Matilainen [Mon, 5 May 2008 06:51:14 +0000 (09:51 +0300)]
Start cleaning up the rpmerr removal mess of build/ return codes
- define PART_ERROR for error returns from various parse bits,
teach parseSpec() about it
- streamline parseSpec() exists
Jindrich Novy [Mon, 5 May 2008 06:12:04 +0000 (08:12 +0200)]
Remove obsolete parseForRegexLang()
- no use for it since _langpatt is now gone
Jindrich Novy [Mon, 5 May 2008 06:05:50 +0000 (08:05 +0200)]
macros.in cleanup
- really deprecate _langpatt
- obsolete _package_version
- use paths more conforming to FHS
- fix typos
Panu Matilainen [Mon, 5 May 2008 05:38:57 +0000 (08:38 +0300)]
Use fprintf() for writing PGP passphrase (like we do for GPG)
- avoids having two separate write()'s to check
- ...if we actually checked the result, gcc doesn't whine about
unchecked fprintf() return...
Panu Matilainen [Mon, 5 May 2008 05:27:07 +0000 (08:27 +0300)]
Check for pipe() returns on GPG and PGP signature creation
Panu Matilainen [Mon, 5 May 2008 05:23:24 +0000 (08:23 +0300)]
Check for pipe() returns in rpmfc helper output reading
Panu Matilainen [Mon, 5 May 2008 05:15:17 +0000 (08:15 +0300)]
Silence bogus warning on memset()
- gcc doesn't realize shrank == 0 would terminate the process via error()
before reaching memset(), work around...
Panu Matilainen [Mon, 5 May 2008 04:37:50 +0000 (07:37 +0300)]
Check for pipe() return code on --pipe
Panu Matilainen [Sat, 3 May 2008 09:44:16 +0000 (12:44 +0300)]
Add missing include
Panu Matilainen [Sat, 3 May 2008 09:43:37 +0000 (12:43 +0300)]
Initialize nb before use
Panu Matilainen [Sat, 3 May 2008 09:38:02 +0000 (12:38 +0300)]
Create signature target rpm in same directory as original
- rename() doesn't work across devices, oops...
Panu Matilainen [Sat, 3 May 2008 09:34:19 +0000 (12:34 +0300)]
Temp file handling tweaks
- rename rpmMkTemp() (back) to rpmMkTempFile()
- rpmMkTemp() is now a lower level thin wrapper around mkstemp()
Panu Matilainen [Sat, 3 May 2008 07:36:00 +0000 (10:36 +0300)]
rpmReSign() doesn't return rpmRC codes, make it more obvious
- cleanup "rc = RPMRC_FAIL" copy-paste leftovers
Panu Matilainen [Sat, 3 May 2008 07:27:16 +0000 (10:27 +0300)]
Eliminate static sized string buffer frpm rpmReSign()
- use rpmMkTemp() for creating the temp file instead of insecure
close(mkstemp())
Panu Matilainen [Fri, 2 May 2008 08:48:27 +0000 (11:48 +0300)]
Check for arch compatibility on multilib too
Panu Matilainen [Fri, 2 May 2008 08:40:29 +0000 (11:40 +0300)]
Ensure default SIGPIPE handler for --pipe (rhbz#444389)
Panu Matilainen [Wed, 30 Apr 2008 11:38:09 +0000 (14:38 +0300)]
RIP alloca()
Jindrich Novy [Wed, 30 Apr 2008 11:24:51 +0000 (13:24 +0200)]
Rewrite xmlFormat() and remove useless xmlstrlen() and xmlstrcpy()
Panu Matilainen [Wed, 30 Apr 2008 10:34:28 +0000 (13:34 +0300)]
Kill of alloca() from tufdio tests
- yes it leaks, who cares
Panu Matilainen [Wed, 30 Apr 2008 10:30:54 +0000 (13:30 +0300)]
Lazily copy basenames in relocation if needed
- couple of tiny array helper funcs to push stuff out of relocateFileList()
instead of adding yet more gunk
- use the helpers for dirnames too
Panu Matilainen [Wed, 30 Apr 2008 09:45:16 +0000 (12:45 +0300)]
Simplify relocation directory name manipulation
Jindrich Novy [Wed, 30 Apr 2008 09:39:49 +0000 (11:39 +0200)]
Don't use alloca in rpmal.c
Panu Matilainen [Wed, 30 Apr 2008 09:29:40 +0000 (12:29 +0300)]
Allocate relocation dirnames on heap, not stack
Jindrich Novy [Wed, 30 Apr 2008 08:32:33 +0000 (10:32 +0200)]
Don't use alloca() in fpsCompare()
Panu Matilainen [Wed, 30 Apr 2008 07:32:57 +0000 (10:32 +0300)]
Begin untangling relocation alloca()'s
- allocate rpmRelocation paths on heap, not stack
Panu Matilainen [Wed, 30 Apr 2008 07:05:48 +0000 (10:05 +0300)]
Kill the remaining alloca in fp doLookup()
Panu Matilainen [Wed, 30 Apr 2008 06:47:25 +0000 (09:47 +0300)]
One alloca() down in fp doLookup()
- single point of exit from the point where buf gets allocated
Panu Matilainen [Tue, 29 Apr 2008 14:38:21 +0000 (17:38 +0300)]
Plug a memleak in lzclose()
Panu Matilainen [Tue, 29 Apr 2008 14:09:07 +0000 (17:09 +0300)]
Plug memleak in parseSimplePart()
Panu Matilainen [Tue, 29 Apr 2008 14:04:08 +0000 (17:04 +0300)]
Plug memleak in macro doFoo()
Jindrich Novy [Tue, 29 Apr 2008 14:08:19 +0000 (16:08 +0200)]
Don't use alloca in rpmGenPath()
Jindrich Novy [Tue, 29 Apr 2008 13:40:20 +0000 (15:40 +0200)]
Calm down compiler warnings due to fsm->patch being const
Jindrich Novy [Tue, 29 Apr 2008 13:01:53 +0000 (15:01 +0200)]
Merge branch 'master' of ssh://login.linux.duke.edu/home/groups/rpm/git/rpm
Jindrich Novy [Tue, 29 Apr 2008 13:01:05 +0000 (15:01 +0200)]
Avoid general ETXBSY and allow it for HP-UX exclusively
- for more info: http://www.uwsg.iu.edu/hypermail/linux/kernel/9808.3/0883.html
- remove the rest of alloca()s
Jindrich Novy [Tue, 29 Apr 2008 12:59:18 +0000 (14:59 +0200)]
Remove useless alloca_strdup()
- and yet more alloca usage in writeFile()
Panu Matilainen [Tue, 29 Apr 2008 12:50:43 +0000 (15:50 +0300)]
Remove unused alloca_strdup() define
Panu Matilainen [Tue, 29 Apr 2008 12:08:47 +0000 (15:08 +0300)]
Allocate directory names on heap in compressFilelist()
Panu Matilainen [Tue, 29 Apr 2008 11:44:28 +0000 (14:44 +0300)]
Use rpmGetPath() to construct full dbpath in db3open()
- another alloca() bites the dust...
Panu Matilainen [Tue, 29 Apr 2008 11:17:09 +0000 (14:17 +0300)]
Oops, one more alloca() in rpmtsOrder()
Panu Matilainen [Tue, 29 Apr 2008 11:13:59 +0000 (14:13 +0300)]
Remove alloca() from rpmtsOrder()
Panu Matilainen [Tue, 29 Apr 2008 11:02:06 +0000 (14:02 +0300)]
Dead code elimination
Panu Matilainen [Tue, 29 Apr 2008 10:49:53 +0000 (13:49 +0300)]
Remove alloca() from rpmvercmp()
- use variable array instead, versions had better be short enough to fit
on stack safely
Panu Matilainen [Tue, 29 Apr 2008 10:33:09 +0000 (13:33 +0300)]
Avoid alloca() in headerLoad()
- REGION_TAG_COUNT is a constant so local array would do even without
C99..
Panu Matilainen [Tue, 29 Apr 2008 10:17:33 +0000 (13:17 +0300)]
Avoid alloca() in findTag()
- instead of putting RPMTAG_ into strings where it's not present,
do it the other way around
- avoids a string copy and redundant string comparisons on the RPMTAG_ part
Panu Matilainen [Tue, 29 Apr 2008 09:59:20 +0000 (12:59 +0300)]
Avoid alloca() in shescapeFormat()
Panu Matilainen [Tue, 29 Apr 2008 09:45:51 +0000 (12:45 +0300)]
Avoid alloca() in headerAddI18NString()
Panu Matilainen [Tue, 29 Apr 2008 09:40:03 +0000 (12:40 +0300)]
This code has been commented out since -99, time to let go...
Panu Matilainen [Tue, 29 Apr 2008 08:54:44 +0000 (11:54 +0300)]
Eliminate remaining alloca() in rpmAddSignature()
Panu Matilainen [Tue, 29 Apr 2008 08:53:13 +0000 (11:53 +0300)]
Plug memleaks in signature generation
Panu Matilainen [Tue, 29 Apr 2008 08:35:30 +0000 (11:35 +0300)]
Avoid alloca() in rpmtsSetRootDir()
Panu Matilainen [Tue, 29 Apr 2008 08:26:00 +0000 (11:26 +0300)]
Avoid alloca() in rpmDetectPGPVersion()
- the whole detection is hysterical and could/should probably be killed
but what the heck...
Panu Matilainen [Tue, 29 Apr 2008 08:24:25 +0000 (11:24 +0300)]
Avoid alloca() in makeGPGSignature()
Panu Matilainen [Tue, 29 Apr 2008 08:16:46 +0000 (11:16 +0300)]
Avoid unchecked stpcpy() to static sized buf from alloca()
Panu Matilainen [Tue, 29 Apr 2008 08:12:15 +0000 (11:12 +0300)]
Avoid alloca() in rpmtsInitIterator()
Panu Matilainen [Tue, 29 Apr 2008 08:02:41 +0000 (11:02 +0300)]
Read "canonvendor" macros by default again
- the whole vendor etc stuff needs rethinking etc, but for now this "works"
Panu Matilainen [Tue, 29 Apr 2008 07:40:56 +0000 (10:40 +0300)]
Avoid alloca() in legacyRetrofit()
Panu Matilainen [Tue, 29 Apr 2008 07:34:54 +0000 (10:34 +0300)]
Use rpmioSlurp() for reading in rpmrc's
- have doReadRC() do the opening of files instead of caller
- always expand rc filename on entry, simplifies includes too
- streamline exit paths
Panu Matilainen [Tue, 29 Apr 2008 06:59:56 +0000 (09:59 +0300)]
Use rpmGlob() for expanding rcfiles globs
- avoids static filename buffer and simplifies things quite a bit
Jindrich Novy [Mon, 28 Apr 2008 15:30:25 +0000 (17:30 +0200)]
Rewrite rpmioMkpath()
Panu Matilainen [Mon, 28 Apr 2008 19:25:47 +0000 (22:25 +0300)]
genCpioListAndHeader() indentation fixes, no other changes
Panu Matilainen [Mon, 28 Apr 2008 18:58:26 +0000 (21:58 +0300)]
Allocate buffer for %doc filenames dynamically
Panu Matilainen [Mon, 28 Apr 2008 18:37:47 +0000 (21:37 +0300)]
Remove unnecessary block which just messes up indentation
Panu Matilainen [Mon, 28 Apr 2008 18:35:26 +0000 (21:35 +0300)]
Set processingFailed just once