Panu Matilainen [Mon, 17 Mar 2008 12:27:03 +0000 (14:27 +0200)]
Nuke leftover rollback test program, update translations
Panu Matilainen [Mon, 17 Mar 2008 12:10:11 +0000 (14:10 +0200)]
Remove repackage support to finish off with rollback
The two main classes of rollback/repackage need are
1) Gimme back the previous version, this broke XYZ on my system.
This is better handled by downgrading to the previous version of original
package instead of repackaged garbage. We should maintain real package
(version) history somewhere.
2) Upgrade messed up my configuration.
Instead of repackaging everything we could be far more intelligent
wrt config files, stick them into real version control or at least have
hooks to do so.
RPMCALLBACK_REPACKAGE_* and RPMTRANS_FLAG_REPACKAGE definitions left around
to avoid needlessly breaking everybodys callbacks (for now)
Panu Matilainen [Mon, 17 Mar 2008 09:53:27 +0000 (11:53 +0200)]
Remove rollback support
- not possible to do reliably within rpm
- effort is better wasted on investigating fs-level snapshots, which
is way beyond rpm scope except for hooks to interact with the snapshot
mechanism to communicate beginning/end of transaction and such
Panu Matilainen [Mon, 17 Mar 2008 10:04:39 +0000 (12:04 +0200)]
Remove support for automatic rollback on failure
- doing this reliably from rpm is simply not possible as there's no way
to undo script actions, might as well not pretend we can
- for a feature that's not generally usable it complicates mainline code
way too much
Panu Matilainen [Fri, 14 Mar 2008 12:55:44 +0000 (14:55 +0200)]
Make arch conditional in default queryformat
- gpg-pubkey's dont have arch, argh...
Panu Matilainen [Fri, 14 Mar 2008 12:27:01 +0000 (14:27 +0200)]
Revert RPMTAG_FILEDIGESTS -> FILEMD5's for now
Panu Matilainen [Fri, 14 Mar 2008 12:18:43 +0000 (14:18 +0200)]
Type missing for RPMTAG_HEADERI18NTABLE
Panu Matilainen [Thu, 13 Mar 2008 07:20:24 +0000 (09:20 +0200)]
Don't process objdump sections after "Version References"
autodeps/linux.req expects the 'Version References' section to be
the very last section in objdump -p output and nothing else to follow
it, but at least on Linux/ARM, you can get output a la:
Version References:
required from ld-linux.so.3:
0x0d696914 0x00 06 GLIBC_2.4
required from libacl.so.1:
0x05822450 0x00 05 ACL_1.0
required from libgcc_s.so.1:
0x0b792655 0x00 04 GCC_3.5
required from librt.so.1:
0x0d696914 0x00 03 GLIBC_2.4
required from libc.so.6:
0x0d696914 0x00 02 GLIBC_2.4
private flags = 4000002: [Version4 EABI] [has entry point]
This causes the built package (if the external dep generator is
used) to end up with bogus dependencies.
Issue found by Rabeeh Khoury <rabeeh@marvell.com>
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Panu Matilainen [Wed, 12 Mar 2008 13:34:49 +0000 (15:34 +0200)]
Remove rpmio internal copy of poptParseArgvString() & friends
- the old included version segfaults with funky input and whatnot...
- link rpmio to real popt instead
Panu Matilainen [Wed, 12 Mar 2008 11:20:47 +0000 (13:20 +0200)]
Refcount dbiTags on db open+close, automatically free
- avoids rpmdb_internal include in rpmqv.c and memleak in API clients
- avoids having to export dbiTags(Max) at all
Panu Matilainen [Tue, 11 Mar 2008 09:33:04 +0000 (11:33 +0200)]
Update translations
Panu Matilainen [Tue, 11 Mar 2008 09:32:37 +0000 (11:32 +0200)]
Dust off Finnish translation
Panu Matilainen [Tue, 11 Mar 2008 06:42:34 +0000 (08:42 +0200)]
Updated Polish translation (Piotr Drąg)
Panu Matilainen [Fri, 7 Mar 2008 13:07:06 +0000 (15:07 +0200)]
Remove support for automatically importing pubkeys from keyservers
- causes hideous load on keyservers and autoimporting without user approval
isn't a very good idea, especially as long as imported == trusted
- nothing wrong with supporting hkp lookups on manual --import though...
- curl doesn't know about hkp:// but http on appropriate port works just
fine, lets re-enable it
Panu Matilainen [Fri, 7 Mar 2008 12:05:28 +0000 (14:05 +0200)]
Remove BDB copy from the repository, it doesn't belong there
Panu Matilainen [Fri, 7 Mar 2008 11:34:20 +0000 (13:34 +0200)]
Stop insecurely covering up for lack of snprintf and vsnprintf
- we require C99 anyway
- snprintf is used unconditionally all around the tree
- if absolutely needed for portability, having insecure wrappers littered
over the source is not the best way
Panu Matilainen [Fri, 7 Mar 2008 11:04:46 +0000 (13:04 +0200)]
Drop librpmbuild from rpm.pc
- next to nothing actually uses librpmbuild, avoid dragging it in needlessly
- if needed, add a separate .pc for rpmbuild later
Panu Matilainen [Fri, 7 Mar 2008 10:36:22 +0000 (12:36 +0200)]
Remove unused RPMUSER, GROUP etc from configure.ac
Panu Matilainen [Fri, 7 Mar 2008 07:47:51 +0000 (09:47 +0200)]
Log message adjustments
- dont waste translator time by having them translate obscure debug messages
- adjust some suspect looking logging levels
- mark some non-debug messages for translation where missing..
Panu Matilainen [Fri, 7 Mar 2008 06:29:03 +0000 (08:29 +0200)]
Further constants made readonly...
Panu Matilainen [Fri, 7 Mar 2008 06:21:55 +0000 (08:21 +0200)]
Make _cacheDependsRC static, not needed outside depends.c
Panu Matilainen [Fri, 7 Mar 2008 06:20:03 +0000 (08:20 +0200)]
Another pile of read-only constants marked as such..
Panu Matilainen [Fri, 7 Mar 2008 06:15:03 +0000 (08:15 +0200)]
Mark another pile of header_magic duplicates as read-only
Panu Matilainen [Thu, 6 Mar 2008 18:47:17 +0000 (20:47 +0200)]
Make yet more build tables readonly
- preambleList and partList are modified runtime for silly reasons,
reminders to fix later..
Panu Matilainen [Thu, 6 Mar 2008 18:25:43 +0000 (20:25 +0200)]
Make headerCopyTags() tag list const, the list isn't modified..
Panu Matilainen [Thu, 6 Mar 2008 18:17:20 +0000 (20:17 +0200)]
Mark bunch of static build-related tables as read-only
Panu Matilainen [Thu, 6 Mar 2008 17:57:01 +0000 (19:57 +0200)]
Make rpmfc token and apply tables readonly
Panu Matilainen [Thu, 6 Mar 2008 17:51:06 +0000 (19:51 +0200)]
Mark rpm tag tables read-only as far as we can..
Panu Matilainen [Thu, 6 Mar 2008 17:50:41 +0000 (19:50 +0200)]
Mark various header tables and definitions as read-only
Panu Matilainen [Thu, 6 Mar 2008 17:49:08 +0000 (19:49 +0200)]
Make the rpmdb vectors readonly
Panu Matilainen [Thu, 6 Mar 2008 17:27:19 +0000 (19:27 +0200)]
Mark bunch of misc. tables & such read-only in librpmio
Panu Matilainen [Thu, 6 Mar 2008 17:25:46 +0000 (19:25 +0200)]
Make rpmio vectors read-only
Panu Matilainen [Thu, 6 Mar 2008 17:24:56 +0000 (19:24 +0200)]
Further read-only tables in rpmpgp..
Panu Matilainen [Thu, 6 Mar 2008 16:56:38 +0000 (18:56 +0200)]
Mark various pgp value tables read-only as they should be
Panu Matilainen [Tue, 4 Mar 2008 07:26:27 +0000 (09:26 +0200)]
PreReq -> Requires in ordering debug output (Pascal Rigaux)
Panu Matilainen [Tue, 4 Mar 2008 06:16:02 +0000 (08:16 +0200)]
Add super-H arch support (rhbz#432496)
Patch from Masaki Chikama
Panu Matilainen [Thu, 28 Feb 2008 10:46:45 +0000 (12:46 +0200)]
Add missing leading underscores to names passed to rpmMkdirPath()
- mostly cosmetical, leading to slightly misleading error messages
Panu Matilainen [Tue, 26 Feb 2008 20:42:46 +0000 (22:42 +0200)]
Another unnecessarily exported internal variable..
Panu Matilainen [Tue, 26 Feb 2008 20:24:28 +0000 (22:24 +0200)]
Make bunch of macro internal variables static
Panu Matilainen [Tue, 26 Feb 2008 20:04:17 +0000 (22:04 +0200)]
Remove hidden --nolibio cli switch, avoid exporting noLibio
Panu Matilainen [Tue, 26 Feb 2008 19:46:38 +0000 (21:46 +0200)]
Un-inline fd*Digest()
- avoid leaking nss + digest internals all over the place
Panu Matilainen [Tue, 26 Feb 2008 19:02:21 +0000 (21:02 +0200)]
Make the rest of rpmio vectors static inside rpmio
- no external needs, gets them out of exported library symbols
Panu Matilainen [Tue, 26 Feb 2008 18:53:38 +0000 (20:53 +0200)]
Remove the last remains of unused and unnecesary rpmio rpc vectors
Panu Matilainen [Mon, 25 Feb 2008 18:16:43 +0000 (20:16 +0200)]
Eliminate rpmdb refcount debug junk from the API
Panu Matilainen [Mon, 25 Feb 2008 10:33:19 +0000 (12:33 +0200)]
Remove some ancient dead code
Panu Matilainen [Mon, 25 Feb 2008 10:24:22 +0000 (12:24 +0200)]
Add %{sources} and %{patches} macros to build
- push full patch + source paths to global lua tables
- %{sources} and %{patches} are just easy-access wrappers to accessing
the lua tables
- "git-am -p1 %{patches}" and enjoy...
Panu Matilainen [Mon, 25 Feb 2008 09:03:38 +0000 (11:03 +0200)]
Enable variadic checks for rpmExpand(), rpmlog() and rpmGetPath()
Suggested by Michal Marek
Panu Matilainen [Mon, 25 Feb 2008 09:02:07 +0000 (11:02 +0200)]
Add new header for collecting utility macros
- atm consists of portability wrappers to various gcc extensions like
__attribute__() to provide additional warnings and checks
- copied from glib to avoid an extra build dependency just for the silly
macros, renamed to avoid namespace clashing
Panu Matilainen [Fri, 22 Feb 2008 13:39:24 +0000 (15:39 +0200)]
Remove ancient commentary from header.h
- history is interesting but public headers aren't the place for it
- "implemented soon after" from 2001 comments are hardly useful :)
Panu Matilainen [Fri, 22 Feb 2008 13:37:16 +0000 (15:37 +0200)]
Remove useless ifdeffed typedefs
- rpmdb & iterator are typedeffed in rpmtypes.h already..
Panu Matilainen [Fri, 22 Feb 2008 13:30:51 +0000 (15:30 +0200)]
Rip ancient, unused setperms + unsetgit make targets
Panu Matilainen [Fri, 22 Feb 2008 13:29:04 +0000 (15:29 +0200)]
Use %__tar instead of hardcoded tar in build prep
Panu Matilainen [Fri, 22 Feb 2008 13:08:41 +0000 (15:08 +0200)]
Don't use legacy gzipbin etc macros in build
Panu Matilainen [Fri, 22 Feb 2008 13:05:16 +0000 (15:05 +0200)]
Make %patch use %__patch instead of hardcoded patch.. patchy huh?
Panu Matilainen [Fri, 15 Feb 2008 13:21:22 +0000 (15:21 +0200)]
Shut up another case of FUSE mount stat whining (rhbz#432907)
Panu Matilainen [Thu, 14 Feb 2008 13:02:18 +0000 (15:02 +0200)]
Improve header i18n locale matching
- Get best lang from rpm HEADERI18NTABLE, instead of getting first fuzzy match
(eg: zh_TW matches zh_CN whereas zh_TW entry is available)
Patch from Pascal Rigaux
Panu Matilainen [Thu, 14 Feb 2008 11:10:51 +0000 (13:10 +0200)]
Fix a bunch of string format mismatches found by gcc's variadic checks
Patch from Michal Marek
Panu Matilainen [Mon, 11 Feb 2008 18:47:03 +0000 (20:47 +0200)]
Automated merge with ssh://pmatilai@login.linux.duke.edu//home/groups/rpm/hg/rpm/
Panu Matilainen [Mon, 11 Feb 2008 18:45:15 +0000 (20:45 +0200)]
Play nice with xargs
- xargs considers exit status 255 special, avoid it. The exit status is
unrealiable anyway so 254 vs 255 shouldn't matter
Patch from Michal Marek
Panu Matilainen [Mon, 11 Feb 2008 18:12:32 +0000 (20:12 +0200)]
Updated sr translation + new sr@latin version from Miloš Komarčević
Peter Jones [Fri, 8 Feb 2008 19:58:26 +0000 (14:58 -0500)]
Don't add syntax errors while I add sorting...
Peter Jones [Fri, 8 Feb 2008 19:54:29 +0000 (14:54 -0500)]
make gendiff sort the file list, so comparing the results is easier.
Panu Matilainen [Fri, 8 Feb 2008 11:45:56 +0000 (13:45 +0200)]
Oops, package type in lead was reversed...
Panu Matilainen [Tue, 5 Feb 2008 15:42:19 +0000 (17:42 +0200)]
Mass convert (back) to rpmTag as it's usable everywhere now
- cast away a few cases where the enum usage causes ridiculous amount
of compiler warnings from unhandled switch-cases
Panu Matilainen [Tue, 5 Feb 2008 14:35:44 +0000 (16:35 +0200)]
Use rpmSigTag enum for signature tags everywhere, more or less..
Panu Matilainen [Tue, 5 Feb 2008 14:14:34 +0000 (16:14 +0200)]
Use rpmTagType always now that it's available everywhere
- RPM_MASK_TYPE is not a datatype, move it out of the enum
- add missing cases python header handling to shut up gcc
Panu Matilainen [Tue, 5 Feb 2008 13:32:33 +0000 (15:32 +0200)]
Couple of missed type abstractions
Panu Matilainen [Tue, 5 Feb 2008 13:29:37 +0000 (15:29 +0200)]
Yet another typedef for transaction ID, use where spotted
- most remaining [u]int16|32 uses are indexes and such, ie actual numbers
Panu Matilainen [Tue, 5 Feb 2008 13:15:41 +0000 (15:15 +0200)]
Use rpm_flag_t for header generic bitfields (file flags etc)
- additionally use rpmfileAtts / rpmVerifyFlags consistently where
appropriate
Panu Matilainen [Tue, 5 Feb 2008 12:50:42 +0000 (14:50 +0200)]
rpm5.org: free 5 bits for rpmfc coloring expansion
- also fixes an overlap between RPMFC_LIBTOOL and RPMFC_MONO which we
had and were out of bits already
Panu Matilainen [Tue, 5 Feb 2008 12:28:06 +0000 (14:28 +0200)]
Couple of int -> rpmVSFlags type corrections
Panu Matilainen [Tue, 5 Feb 2008 12:25:16 +0000 (14:25 +0200)]
Doxygen bits for rpmtypes, much to do still
Panu Matilainen [Tue, 5 Feb 2008 11:53:19 +0000 (13:53 +0200)]
Adjust potfiles from rpmlibprov removal, update...
Panu Matilainen [Tue, 5 Feb 2008 11:52:30 +0000 (13:52 +0200)]
A missed color type in rpmds header
Panu Matilainen [Tue, 5 Feb 2008 11:49:13 +0000 (13:49 +0200)]
Split non-cli verify bits from rpmcli to separate header
Panu Matilainen [Tue, 5 Feb 2008 11:09:13 +0000 (13:09 +0200)]
More rpm-headerdata types: rpm_dev_t and rpm_ino_t, use where spotted
- both hardwired to 32bit internally
Panu Matilainen [Tue, 5 Feb 2008 10:58:05 +0000 (12:58 +0200)]
Add rpm_rdev_t type for rdev entries
- rdev is 16bit whereas dev is 32bit, eww...
Panu Matilainen [Tue, 5 Feb 2008 10:31:32 +0000 (12:31 +0200)]
Add rpm_mode_t type, use where spotted
- bring the types under some sort of control, previously mixture of
int16_t, uint16_t and unsigned short was used..
Panu Matilainen [Tue, 5 Feb 2008 10:12:31 +0000 (12:12 +0200)]
Add rpm_color_t type for rpm color bits, use where spotted
Panu Matilainen [Tue, 5 Feb 2008 09:43:38 +0000 (11:43 +0200)]
Port some rpmds improvements from rpm5.org
- rpmlib provides as rpmds, lose the special case hackery
- new functions rpmdsRpmlib() and rpmdsSearch() + python bindings for them
Panu Matilainen [Tue, 5 Feb 2008 08:08:44 +0000 (10:08 +0200)]
Rip the horrible (ia64) autorelocate kludgery. RIP.
- we don't support relocated x86 on ppc in qemu either...
Panu Matilainen [Tue, 5 Feb 2008 07:41:11 +0000 (09:41 +0200)]
Use helper function to eliminate some copy-paste code
Panu Matilainen [Mon, 4 Feb 2008 10:33:02 +0000 (12:33 +0200)]
Remove double documention for exported header functions
Panu Matilainen [Mon, 4 Feb 2008 10:28:02 +0000 (12:28 +0200)]
Rip the useless header method vectors
Panu Matilainen [Mon, 4 Feb 2008 10:00:30 +0000 (12:00 +0200)]
Make check-programs build again, assorted type-fiddlings..
Panu Matilainen [Mon, 4 Feb 2008 09:50:39 +0000 (11:50 +0200)]
Add rpm_time_t type for time types (in headers), use where spotted
- easy to spot, easy to change...
- time_t size can vary, header data needs fixed size
Panu Matilainen [Mon, 4 Feb 2008 09:27:00 +0000 (11:27 +0200)]
Add rpm_off_t type for file size types, use where spotted
- preliminaries for bumping up the max size, use rpm-specific type
as off_t size varies, header data needs fixed size
Panu Matilainen [Sun, 3 Feb 2008 14:24:07 +0000 (16:24 +0200)]
Make rpmtag.h standalone, include from rpmtypes.h
- permits header.h to use the tag types
- revert other includes back to header.h, more crazy #include churn...
- rpmTagTable etc in header.h for now, don't really belong there but...
Panu Matilainen [Sun, 3 Feb 2008 14:09:20 +0000 (16:09 +0200)]
Use rpmTag etc enums for rpmTagGet*() functions
Panu Matilainen [Sun, 3 Feb 2008 13:41:52 +0000 (15:41 +0200)]
Stuff signature tags to rpmtag.h along with other tags, add enum typedef
Panu Matilainen [Fri, 1 Feb 2008 18:28:31 +0000 (20:28 +0200)]
Fix platform dir creation, doh
Panu Matilainen [Fri, 1 Feb 2008 18:23:25 +0000 (20:23 +0200)]
Update translations for a change..
Panu Matilainen [Fri, 1 Feb 2008 18:18:23 +0000 (20:18 +0200)]
Move platform specific macros under RPMCONFIGDIR/platform/
- reduce the clutter in RPMCONFIGDIR, nicer for packaging too..
Panu Matilainen [Fri, 1 Feb 2008 17:53:04 +0000 (19:53 +0200)]
Ditch the "system" %_topdir, default to ~/rpmbuild instead
- building as root should be discouraged and /usr/src for building stuff
is horribly wrong anyway
Panu Matilainen [Fri, 1 Feb 2008 15:27:22 +0000 (17:27 +0200)]
Don't close GPG/PGP stdin to avoid harmless warning (rhbz#197602)
- patch from rpm5.org
Panu Matilainen [Fri, 1 Feb 2008 11:25:18 +0000 (13:25 +0200)]
Avoid silly fixed size buffer in urlhelper
Panu Matilainen [Fri, 1 Feb 2008 11:00:40 +0000 (13:00 +0200)]
Allocate macro buffers dynamically
- avoids some entirely needless overflow cases
- somewhat similarly to rpm5.org but use heap instead of stack
Panu Matilainen [Fri, 1 Feb 2008 10:11:03 +0000 (12:11 +0200)]
Allocate various macro buffers from heap not stack
Panu Matilainen [Fri, 1 Feb 2008 09:36:13 +0000 (11:36 +0200)]
Define + use MACROBUFSIZ for macro expansion buffer size
- double the buffer size while at it...
Panu Matilainen [Fri, 1 Feb 2008 09:09:13 +0000 (11:09 +0200)]
Macro shell expansion fixes (rhbz#431009)