Panu Matilainen [Wed, 9 Jan 2013 11:48:40 +0000 (13:48 +0200)]
Revert "make "canonicalization(...)shrank by one character" error message more usefull"
- The message is not about basedir and destdir, and printing them
here only makes the message even more confusing than it was, eg:
canonicalization unexpectedly shrank by one character \
('/home/pmatilai/rpmbuild/BUILD/' vs '/usr/src/debug/')
- This reverts commit
1eea433d9961d74dbc797cea7f9bfceab8c16a3e
(cherry picked from commit
e83cc97792d66f8d31ef471455a9fa6ae136ae9a)
Panu Matilainen [Tue, 8 Jan 2013 10:51:36 +0000 (12:51 +0200)]
Alternatively accept files from command line arguments in elfdeps
- Nothing actually needs this, but for testing & debugging command line
args are often nicer than having to pass stuff via stdin.
(cherry picked from commit
6941d51e7e5139014418d1056acb6dafccefbab7)
Panu Matilainen [Mon, 7 Jan 2013 12:23:13 +0000 (14:23 +0200)]
Fake ELF sonames by default (again)
- (Private) libraries which might intentionally not have DT_SONAME
are still recorded as requires from DT_NEEDED, and there's no
way of knowing what's an internal library when generating requires.
Not faking the soname in these cases will only result in broken
requires in cases that always used to "just work".
- Change the switch to --no-fake-soname disabler instead to allow
tweaking in special cases but by default we gotta match linker
(and ELF specification) behavior, no matter how much we'd like to
use this for our own heuristics :-/
(cherry picked from commit
b6f159a34c946b75cb95a7862968ff4e8eb74fe6)
Panu Matilainen [Thu, 3 Jan 2013 15:01:42 +0000 (17:01 +0200)]
Centralize elf soname provide addition, comment
- Remember DT_SONAME in the elfInfo struct if encountered and
only add it after everything else has been processed. This doesn't
change any actual functionality for now, but gives us a single
place where to control the addition.
- Additionally document what the related DT_DEBUG test is for and
clean up the processDynamic() loop and switch-case a bit.
(cherry picked from commit
8e3ce7417a20dd8bca4676c0a598fc4932d22eef)
Panu Matilainen [Thu, 3 Jan 2013 12:16:20 +0000 (14:16 +0200)]
Only fake ELF soname's if requested
- This helps cutting down the number of bogus provides from dlopen()'ed
plugins and internal libraries which preferrably shouldn't have a
soname at all. Unfortunately libtool always puts in a soname even if
-module -avoid-version is used :-/
- OTOH there are broken libraries which dont have a soname even though
they should, so (we need to) allow falling back to the former behavior
of faking up a soname from the basename.
(cherry picked from commit
6c0bdc393dda3de8bacd3a1844d980802e45d195)
Panu Matilainen [Thu, 3 Jan 2013 10:59:13 +0000 (12:59 +0200)]
Move elf class marker figuring to helper function, generalize a bit
- Determine arch-specific issues by looking at the elf header instead
of compile-time #ifdefs, so we'll generate correct dependencies for
non-native elf binaries too. Currently this is just alpha which despite
being a 64bit system, never had the (64bit) markers in its dependencies.
Of course alpha has pretty much already met its mark^H^H er maker by now,
but doing the right thing is cheap... Also we'll need similar special
cases sooner or later for other archs (such as x32).
(cherry picked from commit
f84a71cdc786a012fc9cbe16c5236c622ea970a4)
Panu Matilainen [Thu, 3 Jan 2013 10:11:26 +0000 (12:11 +0200)]
Consolidate elf dependency string generation to helper function
- Figure out the elf class marker early and handle the rest in helper
to avoid copy-slop code. No functional changes.
(cherry picked from commit
e7489abd6672e8c980d2d1b2ab32b963e1ea83b2)
Panu Matilainen [Wed, 19 Dec 2012 11:21:58 +0000 (13:21 +0200)]
Disable useless empty translations, enable Ukrainian translation
(cherry picked from commit
8774cb03cf78b02411cc26329a3c362d95e36fe4)
Thierry Vignaud [Wed, 19 Dec 2012 10:52:51 +0000 (12:52 +0200)]
Initial Breton (br) translation
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
(cherry picked from commit
d277537bcafa65f1cb09d12e259569d6ed35e2a2)
Panu Matilainen [Wed, 19 Dec 2012 10:32:52 +0000 (12:32 +0200)]
Handle transaction global pool allocation centrally in rpmtsPool()
- Previously the pool would only get allocated on successfull addition
of install or erasure elements, causing assert() failures on
operations on empty transaction set, which should be just a no-op,
not an error.
- Make rpmtsPool() create the pool if it doesn't exist, update relevant
users to call rpmtsPool() instead of directly accessing tsmem->pool,
this avoids having to worry about pool existence in all the various cases.
- Also fix up the pool-related comment on rpmtsEmpty(): pools does not
and can not support emptying as it could break references to its
contents. Per-string refcount would be needed for emptying support.
(cherry picked from commit
d73535e1a9dc5095e78475adc5b636d99f01efa9)
Panu Matilainen [Tue, 18 Dec 2012 11:22:12 +0000 (13:22 +0200)]
Clean up fix grabData() error handling in callers
- grabData() only returns non-NULL on data length > 0, separately
checking for length only makes it look like a possible case when
its not. If it were, it'd be a memory leak.
(cherry picked from commit
aa6af71d457be5d78252f23ea68322dcaf199e4e)
Thierry Vignaud [Mon, 17 Dec 2012 12:58:44 +0000 (13:58 +0100)]
display the actual unknown option instead of '?'
'?' is returned by getopt when option is unknown, making hard to
pinpoint the actual bogus option...
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
(cherry picked from commit
8192e580a0a962b8eb7f9d8fc07bd3f99ccf0ffe)
Panu Matilainen [Fri, 14 Dec 2012 07:26:02 +0000 (09:26 +0200)]
Dead code removal
- Nothing is setting "arch" here anymore, this should've been removed in
commit
fddfee17c3c4e19269665800d557d3452cbde162 already
(cherry picked from commit
1c3f966903743ea571a23e1af08a618b4179c5cb)
Panu Matilainen [Mon, 10 Dec 2012 13:20:51 +0000 (15:20 +0200)]
Preparing for 4.11.0-beta1
Panu Matilainen [Fri, 7 Dec 2012 11:54:23 +0000 (13:54 +0200)]
Ensure correct return code on malformed signature in packages
- rpmpkgRead() starts with assumed failure, but there are a number
of places assigning the return code, and by the time we get
to the parsePGPSig() calls its likely to be RPMRC_OK, so the
jumps to exit result in "all is well" return code on a signature
we couldn't even parse. Oops.
- Set the failure status explicitly to fix this fairly nasty regression
introduced in commit
e8bc3ff5d780f4ee6656c24464402723e5fb04f4, ie
rpm >= 4.10.
(cherry picked from commit
96a616c6aed4c516789a154af188f005caf23f14)
Panu Matilainen [Tue, 4 Dec 2012 12:06:42 +0000 (14:06 +0200)]
Handle allocation failure in NSS DSA signature initialization
(cherry picked from commit
2d39f13064f6e6ff1a1633b708e7ba5622a7e8eb)
Panu Matilainen [Fri, 30 Nov 2012 10:23:42 +0000 (12:23 +0200)]
Fix build with older NSS versions that dont support SHA224
- SHA224-support was added around NSS 3.13, dont break compilation
with older versions just for this rarely used hash.
- HASH_AlgSHA224 is an enum so test for SHA224_LENGTH define instead
(cherry picked from commit
e15e29823451663e27a5bec6c865b37901f1c315)
Panu Matilainen [Fri, 30 Nov 2012 10:20:05 +0000 (12:20 +0200)]
Condence NSS getHashType() a bit, handle SHA224 too
- Not much point in enumurating unsupported cases, or with switch-breaks
after returns.
- Recognize SHA224 too
(cherry picked from commit
466a6b554e7f7ded9f1c59646e6b64243ce16812)
Panu Matilainen [Fri, 30 Nov 2012 06:15:24 +0000 (08:15 +0200)]
Unbreak --setperms (RhBug:881835)
- Fixes a regression introduced in rpm >= 4.10 caused by query format
simplification (loss of zero padding support), in commit
1f1e5e88a1f6c60cf2d11014d809000e97a63aad.
- 'chmod' command doesn't need zero padding, just remove the formatting.
In fact we shouldn't be passing the entire mode to it but just the
permission bits, but fortunately chmod isn't too picky here.
(cherry picked from commit
4c90c79e3f3c178189ad3bd89dfd87cd3a2290ab)
Panu Matilainen [Thu, 29 Nov 2012 10:35:13 +0000 (12:35 +0200)]
Get rid of long since deprecated VFY_VerifyDigest() uses
- VFY_VerifyDigest() has been deprecated since NSS >= 3.12 and for
a good reason too: with VFY_VerifyDigest() caller needs to painfully
enumerate every possible supported enc + hash combination, only for
NSS to revert the process. Use the saner VFY_VerifyDigestDirect()
interface instead and test for its presence in configure.
- This means we now require NSS >= 3.12 but as that's already 4.5 years
old and included in ancient beasts like RHEL-4, this doesn't seem
exactly unreasonable requirement. And then there's always beecrypt...
(cherry picked from commit
9b995a7674adba08248fac79ae8b23ecbecc13de)
Panu Matilainen [Thu, 29 Nov 2012 10:31:52 +0000 (12:31 +0200)]
Remove extra ',' in NSS configure checks, whoops
(cherry picked from commit
e1c67173d38016a88989f5b09e0934f90094fe85)
Panu Matilainen [Thu, 29 Nov 2012 09:05:04 +0000 (11:05 +0200)]
Use NSS-defined constants for DSA q-bits and signature length too
- No functional changes here (either), but using the proper constants
simply makes things that little bit clearer.
(cherry picked from commit
604da18bc267f89568081c1ab163c7fc45207220)
Panu Matilainen [Thu, 29 Nov 2012 08:43:24 +0000 (10:43 +0200)]
Avoid deprecation warnings from NSS >= 3.14
- NSS >= 3.14 introduces support for DSA2 and marks DSA_SUBPRIME_LEN
as deprecated. Use explicit DSA1_SUBPRIME_LEN (we only support DSA1
for now) instead where available, add compatibility define for
older versions.
- Also directly include <blapit.h> where its defined - blabit.h gets
included via cryptohi.h but being explicit about it avoids having
to redefine it again "just in case".
(cherry picked from commit
73be0fc7d9faab9f1120e2979f2ca1feed522195)
Panu Matilainen [Mon, 26 Nov 2012 11:34:06 +0000 (13:34 +0200)]
Add missing <errno.h> include
- Should've been in commit
cbd6ef58bbc122e6adf2138679915bd3845d6756,
this breaks build when selinux-support is not compiled in
(cherry picked from commit
ebba1c7b58ba089bb9e1ded147c789d9ca067b6c)
Panu Matilainen [Thu, 22 Nov 2012 10:22:06 +0000 (12:22 +0200)]
Avoid locale issues in rpm2cpio.sh (RhBug:878363)
- In gawk >= 4.0.x printf() "adjusts" things according to current locale,
but we need the data as it is. Forcing C locale prevents gawk from
getting funny ideas about character conversions...
(cherry picked from commit
c0e95f1cedefd6d7522ef2cf7a2ada7c83821596)
Panu Matilainen [Sat, 17 Nov 2012 15:57:59 +0000 (17:57 +0200)]
Fix double-free on %caps in spec (RhBug:877512)
- Fixes regression caused by dumb refactoring mistake in commit
807ba93a6cf168410e3801347898949f356fcd6c
(cherry picked from commit
6bdd34c451dbf69792c59704e06f1ccb045ae743)
Panu Matilainen [Fri, 16 Nov 2012 11:21:15 +0000 (13:21 +0200)]
Add --undefine cli switch for undefining macros (related to RhBug:876308)
(cherry picked from commit
644f080a4868208bf61f28103b7b676c094129a0)
Panu Matilainen [Wed, 14 Nov 2012 10:19:49 +0000 (12:19 +0200)]
Account for temporary disk-space requirement on forced replace too
- Similar to commit
85df102165fdbe64978f2019d757d400e7448218, but
for forced file replacements.
(cherry picked from commit
b1d3f21a54eb1f23a08e9e2d943a6438502b0e47)
Panu Matilainen [Wed, 14 Nov 2012 08:54:38 +0000 (10:54 +0200)]
Remove ancient, no longer relevant comment
- Since commit
f7f5f88f9f3d6587e747b034ccb64a3f00ff4e1e, only the first
instance of shared files is created, the rest are FA_SKIP which
consume no disk space and are not counted either.
(cherry picked from commit
f8da4f613680b4da4ee191f8b7451de9b36bfcbd)
Panu Matilainen [Wed, 14 Nov 2012 08:31:15 +0000 (10:31 +0200)]
Account for temporary disk-space requirements on updates (ticket #175)
- When updating packages, we first create them with a temporary names
and only after all files from payload have been created this way,
the files are renamed to the final target. This means that performing
an update temporarily requires roughly twice the disk space (and inodes)
compared to the final result on per-package level. Which matters
when space is tight, such as presumably in RhBug:872314.
- Simulate what happens on upgrades by adding block and inode delta
to the equation: installing a file always consumes an inode and
the specified amount of disk space. But when replacing files,
reduce the size-delta from disk consumption *after* checking for
problems in a given DSI.
- Also fixes inode accounting which has been broken for forever (since
commit
a9a1fd866c573f41287e6ad256ce64b3970a1eaa more exactly)
(cherry picked from commit
85df102165fdbe64978f2019d757d400e7448218)
Panu Matilainen [Fri, 9 Nov 2012 06:21:51 +0000 (08:21 +0200)]
Add $RPM_CONFIGDIR/macros.d/ directory to default macro path (RhBug:846679)
- There of course already is a drop-in directory for macro config bits
in $(sysconfdir)/rpm/ but as /etc is supposed to be per-host admin
territory, having another place for system package provided rpm config
doesn't seem unreasonable.
- Whether this early in the path is the best possible or even reasonable
position for this remains to be seen... but this way it could be used
for "factory default" config bits by rpm itself as well, while still
permitting vendor/host/user override.
(cherry picked from commit
dcd261b9fe1e61e8ac3e99dbb443145bbf896886)
Panu Matilainen [Thu, 8 Nov 2012 08:26:18 +0000 (10:26 +0200)]
Add cli-switch to query for license files
(cherry picked from commit
4b19c17f53f56c841903e5d6a09efdb10b2ca400)
Panu Matilainen [Thu, 8 Nov 2012 08:20:24 +0000 (10:20 +0200)]
Erm, actually mark the special %license files as RPMFILE_LICENSE
- Somehow I thought this already got done somewhere... the new special
%license files need to be marked as RPMFILE_LICENSE for easy querying.
(cherry picked from commit
5998b02665c30a5b560f8fe899c19235f9299bd0)
Panu Matilainen [Tue, 6 Nov 2012 11:29:11 +0000 (13:29 +0200)]
Document --define in rpmbuild manual as well (RhBug:665536)
Panu Matilainen [Mon, 5 Nov 2012 15:15:23 +0000 (17:15 +0200)]
Dont complain about non-colored arch-specific packages
- Non-colored arch-specific packages are very common, for example
-devel and -debuginfo packages among several other cases for which
we cant do much about. Dont whine without a good reason.
Panu Matilainen [Mon, 5 Nov 2012 14:00:39 +0000 (16:00 +0200)]
Work around Fedora-specific patch to libselinux causing warnings
- In recent Fedora, struct selinux_opt value member is changed to an
anonymous union of char * and char **. Initialize the members by name
to avoid warnings in both (patched and unpatched) cases.
Dominique Leuenberger [Mon, 5 Nov 2012 13:45:29 +0000 (15:45 +0200)]
Fix strncat() boundaries in Fdopen()
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
Panu Matilainen [Mon, 5 Nov 2012 13:08:28 +0000 (15:08 +0200)]
Update contact address in COPYING (RhBug:742362)
- marc and ewt haven't been valid email addresses in the last 10+
years, cough...
Panu Matilainen [Mon, 5 Nov 2012 08:22:50 +0000 (10:22 +0200)]
Update library current + age (ie libs are now 3.1.0)
- We haven't removed or changed any interfaces in a way that would
require full soname bump, only a handful of new interfaces have
been added.
- There aren't actually any new interfaces in librpmbuild or librpmsign
but for sanity and consistency's sake they're all updated...
Panu Matilainen [Mon, 5 Nov 2012 08:09:18 +0000 (10:09 +0200)]
Update translations from Transifex
Panu Matilainen [Thu, 1 Nov 2012 10:09:30 +0000 (12:09 +0200)]
Enable built-in %dirname macro
- This might've clashed with %dir in spec %files at some point as
indicated by the comments, but that should no longer be the case.
Panu Matilainen [Thu, 1 Nov 2012 09:45:44 +0000 (11:45 +0200)]
Rename -debug to -debuginfo to match what everybody uses anyway
- Practically everybody renames the debug information sub-packages
from -debug to -debuginfo, might as well do so upstream already.
Intended to do this like five years ago but kept forgetting...
thierry.vignaud@gmail.com [Fri, 19 Oct 2012 08:08:35 +0000 (10:08 +0200)]
make "canonicalization(...)shrank by one character" error message more usefull
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
Tim Landscheidt [Wed, 24 Oct 2012 23:18:57 +0000 (23:18 +0000)]
Don't call "rpm --addsign" if rpmbuild didn't create RPMs.
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
Panu Matilainen [Thu, 1 Nov 2012 07:45:15 +0000 (09:45 +0200)]
Hide the strpool-related rpmts/rpmfi/rpmds interfaces out of sight for now
- In the package/transaction related things the strpool is more of
an internal implementation detail than an end-goal in itself, move
string pool related interfaces of rpmts, rpmfi and rpmds to
internal-only APIs for now. The kind interfaces we'll want to eventually
export a) dont exist yet and b) are likely to be something very different.
- The string pool itself remains exported however, its a handy data
structure for all sorts of things and both librpm and librpmbuild
heavily use it already.
Panu Matilainen [Wed, 31 Oct 2012 07:43:38 +0000 (09:43 +0200)]
Use private NSS context if NSS supports it (RhBug:871485)
- Older NSS versions operate on global context, which can cause
all sorts of trouble when an API user tries to use NSS for their
own purposes: eg they might want to use NSS databases which is not
possible once we've initialized NSS with NSS_NoDB_Init(). Further
background on the subject at https://wiki.mozilla.org/NSS_Library_Init
- Use private private NSS context when possible (NSS >= 3.12.5) to
avoid such clashes, but keep support for older versions for now.
Panu Matilainen [Tue, 30 Oct 2012 10:38:20 +0000 (12:38 +0200)]
Handle EINTR in Fread() and Fwrite()
- Not all systems automatically restart signal-interrupted operations,
handle this centrally since its so easy to do. Also related to
RhBug:580974.
Panu Matilainen [Tue, 30 Oct 2012 09:36:56 +0000 (11:36 +0200)]
Handle partial reads and errors in package IO (RhBug:802839, RhBug:580974)
- There are no guarantees Fread() will return all of the requested size:
it can return partial data eg on signals and pipe descriptors. Introduce
a helper function to handle this centrally for all package read IO,
effectively reintroducing timedRead() but without the caveats:
timedRead() did not work on compressed streams, did not handle
eg EINTR correctly and while really being an internal helper,
was exported in the API.
Panu Matilainen [Tue, 30 Oct 2012 07:29:46 +0000 (09:29 +0200)]
Fix missing error on --import on bogus key file (RhBug:869667)
- When the "BEGIN PGP" marker is not found at all, we would silently
exit with success when trying to import utter garbage, such as
rpmkeys --import /bin/bash (not that I consider bash as gargabe ;)
Panu Matilainen [Fri, 26 Oct 2012 09:15:07 +0000 (12:15 +0300)]
Remove transition-time leftover pool unfreezing
- Nobody is freezing our pool so there's no need to unfreeze it either...
Panu Matilainen [Fri, 26 Oct 2012 08:47:31 +0000 (11:47 +0300)]
Eliminate static "misc stuff pool" from rpmfi
- Prior to string pool existence, the static "caches" were kinda
necessary for sharing the relatively static content of user/group
names and file languages, but this is nothing but an unfreeable
block of memory at this point. Just use the same pool as everything
else - whether private or shared.
Panu Matilainen [Fri, 19 Oct 2012 13:20:02 +0000 (16:20 +0300)]
Eliminate copy-paste nibble(), use rnibble() instead
Panu Matilainen [Fri, 19 Oct 2012 12:45:14 +0000 (15:45 +0300)]
Default to NSS crypto though...
- NSS is a big and quirky monster for our needs but that's what we've
been defaulting to for the last years, and then there are those
certifications...
- Also default to external beecrypt even if --with-beecrypt is used
Michael Schroeder [Fri, 19 Oct 2012 12:43:09 +0000 (15:43 +0300)]
Add support for beecrypt as an alternative crypto implementation
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
Panu Matilainen [Fri, 19 Oct 2012 10:54:11 +0000 (13:54 +0300)]
Dont bother with digest and link comparisons if size differs, part II
- If link or regular file sizes differ, they cannot possibly be
identical. Saves us from doing bunch of pointless string and
memory comparisons on what's a very busy path - not exactly a huge
with but measurable nevertheless.
Panu Matilainen [Fri, 19 Oct 2012 10:45:04 +0000 (13:45 +0300)]
Dont bother with %config digests on conflicts if size differs
- Files of different size cannot possibly be identical, dont bother
with expensive digest calculation.
Panu Matilainen [Thu, 11 Oct 2012 14:57:10 +0000 (17:57 +0300)]
Fix noarch __isa_* macro filter in installplatform (RhBug:865436)
- The filter wasn't doing what it was supposed to due to extra single
quotes getting inserted, causing "rpmbuild --target noarch foo.spec"
to whine about empty macro bodies. This is a regression introduced
in rpm 4.10, commit
07ec480c180e4005a629242b8f9f8ab640e3e950 to be
precise.
Panu Matilainen [Thu, 11 Oct 2012 12:14:04 +0000 (15:14 +0300)]
Eliminate now unnecessary WITH_SELINUX conditionals in sepolicy plugin
Panu Matilainen [Thu, 11 Oct 2012 12:09:15 +0000 (15:09 +0300)]
Only try compiling sepolicy plugin if selinux is enabled
Panu Matilainen [Thu, 11 Oct 2012 12:04:54 +0000 (15:04 +0300)]
Missing include in string pool
- When compiled without selinux support, stdlib.h doesn't get
included here. Wtf?
Florian Festi [Thu, 11 Oct 2012 10:20:06 +0000 (12:20 +0200)]
Revert "Update man page to reflect that rpm no longer uses MD5 as a default digest algorithm"
This reverts commit
b040b323d6fac32f7f4fa77fd5cb60c7967fc53f.
The change does only affect file digests.
Florian Festi [Wed, 10 Oct 2012 12:59:32 +0000 (14:59 +0200)]
Add rpm --setperms and rpm --setugids to the man page
Florian Festi [Wed, 10 Oct 2012 11:17:09 +0000 (13:17 +0200)]
Update man page to reflect that rpm no longer uses MD5 as a default digest algorithm
Anders F Bjorklund [Wed, 12 Sep 2012 21:26:38 +0000 (23:26 +0200)]
Add lua 5.2 support.
Add compatibility support for both lua-5.1 and lua-5.2,
assuming that the LUA_COMPAT might have been disabled.
Signed-off-by: Panu Matilainen <pmatilai@redhat.com>
Panu Matilainen [Mon, 8 Oct 2012 06:27:18 +0000 (09:27 +0300)]
Don't bother calculating digests of %ghost %config files
- Calculating digest of %ghost config is useless as there's nothing
to compare the result to. Also we never take backups of %ghost
configuration anyway, so this is a total waste of time. One common
case of %ghost %config is the rpmdb (environment and all) which
can be rather large and calculating digests can take several seconds,
only for the results to be thrown away unused.
- There are some cases where it might be reasonable to back up %ghost
%config (eg if it gets replaced by non-config), but
rpmfiConfigConflictIndex() doesn't have sufficient context to figure
that out. For now, preserve the traditional simple rule: no backups
for ghosts, ever.
Panu Matilainen [Fri, 5 Oct 2012 08:38:26 +0000 (11:38 +0300)]
Use rpmal lookups for already added obsoletions too
- Related to commit
bee348b5d101a0ea6039b56c39e8ddc1cfc09a1a,
use the newly added obsoletes hash to lookup already added
obsoleters. Eliminates the dumb linear lookup and is unsurprisingly
a whole lot faster on larger transactions.
Panu Matilainen [Fri, 5 Oct 2012 08:35:11 +0000 (11:35 +0300)]
Add a hash + lookup function on obsoletes to rpmal
- Similar lazily created hash as provides for fast obsoletes lookups.
This is so similar the provides that creation etc functions should
be unified, but leaving that exercise till later.
Panu Matilainen [Fri, 5 Oct 2012 08:34:22 +0000 (11:34 +0300)]
Rename rpmalProvidesHash to rpmalDepHash
- The same hash type is valid for any ds and we'll be adding more...
Panu Matilainen [Fri, 5 Oct 2012 08:25:49 +0000 (11:25 +0300)]
Avoid changing the file info sets when adding to rpmal file hash
- Use indexed access to the file info sets so we're not mucking
around with somebody elses rpmfi iterator index (currently not an
issue but still...)
Panu Matilainen [Fri, 5 Oct 2012 07:16:44 +0000 (10:16 +0300)]
Avoid changing the dependency sets when looking up from rpmal
- Use indexed access and comparison to avoid mucking around with
eg transaction element dependency sets while the callers are walking
over them.
Panu Matilainen [Fri, 5 Oct 2012 07:13:09 +0000 (10:13 +0300)]
Avoid changing the dependency sets when adding to rpmal hash
- Use indexed access to the dependencies so we're not mucking with
the rpmds iterator index behind anybodys back, this could affect
all sorts of things but miraculously nothing is hitting it atm...
Panu Matilainen [Fri, 5 Oct 2012 07:09:50 +0000 (10:09 +0300)]
Guard against NULL ds in the rpmds indexed getters
- NULL dependency sets are all over the place as NULL ds means
"no dependencies of this kind", whereas for eg rpmfi NULL is an
error.
Panu Matilainen [Fri, 5 Oct 2012 07:07:52 +0000 (10:07 +0300)]
Export the indexed rpmds accessors internally
- Add an internal header for rpmds too to allow adding interfaces we
dont necessarily want to export in the public API, make the indexed
accessors available internally.
Panu Matilainen [Fri, 5 Oct 2012 07:04:04 +0000 (10:04 +0300)]
Add indexed access variants for rpmdsColor() and rpmdsCompare()
- Various places in rpm need random access to the dependency sets,
save and restore on somebody elses "iterator index" just doesn't
cut it. This is merely preliminaries for further changes.
Panu Matilainen [Thu, 4 Oct 2012 12:48:43 +0000 (15:48 +0300)]
Be as lazy as possible wrt rpmal hash creation
- Separate provides and files hash creation, delay both until the last
moment before first valid lookup. In practise, this means the provides
hash is created early due to lookups from rpmtsAddInstallElement(),
but the big bad file hash creation is delayed until the entire
transaction set has been more-or-less populated. Which means we have
a better idea about the necessary hash table size, meaning fewer
hash resizes, resulting in good deal faster execution with no
downsides - if something happened to trigger an early file lookup
it'll all still work, just slower.
Panu Matilainen [Tue, 2 Oct 2012 13:40:45 +0000 (16:40 +0300)]
Use rpmal lookups for finding redundancies in the added set
- Replace the dumb linear search across all elements on each
addition with (filtered) rpmal lookups where possible. rpmal
doesn't (yet) have obsoletes information so for already added
obsoletions we have no choice but to walk the walk.
- As a result, findPos() is hugely faster for large transactions
but rpmal hashes are now generated on the fly for everything
instead of doing it all at once before the actual dependency
checks / ordering, which will cost us something in terms of
hash table resizes.
Panu Matilainen [Tue, 2 Oct 2012 13:17:05 +0000 (16:17 +0300)]
Dont bother with rpmRelocateFileList() if relocations have been specified
- rpmRelocateFileList() doesn't modify anything when no relocations
are to be done, but what it does is not exactly free, unnecessarily
calling it is dumb.
Panu Matilainen [Tue, 2 Oct 2012 09:12:20 +0000 (12:12 +0300)]
Only return non-NULL from rpmalAllSatisfiesDepends() on real matches
- The provides hash lookup can and does return hits that dont actually
satisfy the dependency. Dont bother callers with apparent hits
(ie non-NULL returns) when nothing actually matches the dependency.
Panu Matilainen [Tue, 2 Oct 2012 07:50:10 +0000 (10:50 +0300)]
Export rpmalAllSatisfiesDepend() internally (again)
- For some uses, we need to be more selective about providers...
Panu Matilainen [Tue, 2 Oct 2012 07:37:36 +0000 (10:37 +0300)]
Pass the newly create ts element instead of header to findPos()
- Avoids a little bit of extra work, we already have the relevant
bits of information in the rpmte and grabbing them from there is
cheaper than looking up stuff from headers. Also avoids creating
another copy of the new elements obsoletes dependency set unnecessarily.
Panu Matilainen [Mon, 1 Oct 2012 10:19:59 +0000 (13:19 +0300)]
Eliminate all the now unnecessary fsm->ix save-and-restores
- Now that the relevant places are accepting file index as argument,
we no longer need to save and restore fsm->ix in all the places
dealing with hard links.
Panu Matilainen [Mon, 1 Oct 2012 10:15:30 +0000 (13:15 +0300)]
Pass file index as argument to fsm fsmMapPath()
- Normally this is just the current index, but when writing links
its something else and makes sense as an argument (so we dont
need to save and restore fsm->ix when doing something different)
Panu Matilainen [Mon, 1 Oct 2012 10:01:33 +0000 (13:01 +0300)]
Pass file index as argument to fsm fsmCommit()
- Normally this is just the current index, but when writing links
its something else and makes sense as an argument (so we dont
need to save and restore fsm->ix when doing something different)
Panu Matilainen [Mon, 1 Oct 2012 09:54:15 +0000 (12:54 +0300)]
Pass file index as argument to fsm writeFile()
- Normally this is just the current index, but when writing links
its something else and makes sense as an argument (so we dont
need to save and restore fsm->ix when doing something different)
Panu Matilainen [Mon, 1 Oct 2012 09:30:28 +0000 (12:30 +0300)]
Eliminate current hardlink set from fsm struct
- Return the hard link set from saveHardLink() when ready into a local
variable in the only place that cares: rpmPackageFilesInstall().
Another pointless and hard-to-follow fsm-global state variable
gone...
Panu Matilainen [Mon, 1 Oct 2012 09:16:18 +0000 (12:16 +0300)]
Pinpoint the one place where fsm->li is modified on real purpose
- saveHardLinks() is the only place where fsm->li value matters on
return: during installation its used for "returning" the current
link set when all the links in that set have been seen so they
are ready for creating.
Panu Matilainen [Mon, 1 Oct 2012 09:04:29 +0000 (12:04 +0300)]
Avoid using fsm->li directly when creating pending hard links
- fsmMakeLinks() operates on the current link at hand, doesn't walk
the links or anything... pass the link as an argument from the sole
caller and operate on that.
Panu Matilainen [Mon, 1 Oct 2012 08:53:48 +0000 (11:53 +0300)]
Avoid using/modifying fsm->li when committing hardlinks
- Using a "global" variable for local... more of the same:
fsmCommitLinks() gets called at the end of each round of the install
loop if there are links, it does the actual link discovery on its
own and fsm->li state will get rewritten by saveHardLink() at the
start of the next round.
Panu Matilainen [Mon, 1 Oct 2012 08:47:29 +0000 (11:47 +0300)]
Avoid using/modifying fsm->li when writing links to payload
- Using a "global" variable for local iteration is just... this gets
called once at the end of package build, fsm->li carries no
global state here.
Panu Matilainen [Mon, 1 Oct 2012 08:39:47 +0000 (11:39 +0300)]
Avoid using/modifying fsm->li when freeing hard link sets
- Using a "global" variable for local iteration is...
Panu Matilainen [Mon, 1 Oct 2012 08:37:01 +0000 (11:37 +0300)]
Avoid using/modifying fsm->li on hard link checking
- Using a "global" variable for local iteration is just dumb...
Panu Matilainen [Mon, 1 Oct 2012 08:33:08 +0000 (11:33 +0300)]
Move hardlink saving out of fsmInit()
- Now that we have separate functions for install, erase and build,
handle the hardlink saving locally where it matters.
Panu Matilainen [Mon, 1 Oct 2012 06:54:35 +0000 (09:54 +0300)]
Free hardlink sets centrally in fsmFree()
- Having three places doing the same thing doesn't make a whole
lotta sense...
Panu Matilainen [Mon, 1 Oct 2012 06:23:29 +0000 (09:23 +0300)]
Filter out skipped files on hardlink checking
- Legitimately skipped files (links) must not cause install-errors.
This has always been broken, but the errors were completely ignored
on install prior to rpm 4.10, and now that we're only creating the
first instance of a shared file, secondary arch multilib packages
with hardlinks were causing install failures because of the "missing"
hardlinks here.
- The relevant part here is obviously the XFA_SKIPPING() test, for
which we need the file states. To keep the lines short, grab the
index to a helper variable and replace other fsm->li->filex[i] uses
with that too.
Panu Matilainen [Fri, 28 Sep 2012 09:43:39 +0000 (12:43 +0300)]
Simplify fpLookupSubdir() a bit
- Use string offsets for basename start and end to track the progress,
avoiding +1/-1 adjustments in every damn calculation.
- Reduce the places where new basename is calculated to just one
at the start of the main loop, just adjust the basename start
and end accordingly beforehand.
- This shouldn't change any functionality, just simplify the code
a little bit.
Florian Festi [Fri, 28 Sep 2012 07:23:03 +0000 (10:23 +0300)]
Make string pool strings static in memory
- Use multiple chunks that get allocated as old ones get filled up
instead of reallocating, store direct pointers to the strings in
the id array.
- This prevents nasty surprises when previously retrieved pointer
to a pool string goes invalid underneath us due to somebody
adding a new string, and restores former rpm API behavior:
string pointers retrieved from eg rpmds and rpmfi are valid for
the entire lifetime of those objects.
Panu Matilainen [Thu, 27 Sep 2012 12:56:56 +0000 (15:56 +0300)]
Minimally resurrect fpLookupSubdir() functionality
- fpLookupSubdir() has been broken for a long, long time, probably
because some subtle extra/missing slash issue. It got totally
wrecked by the pool changes though: there are heading and trailing
slashes everywhere and the calculations were off-by-one to every
possible direction because of that. And the previous attempts
(eg commit
566a332c6907a0969ea8f79a4c7a62bca2d1f1b4) makes
even a bigger mess of it.
- This is far messier than it needs to be, but for now going for
minimal resurrection rather than rewrite it all at once. It's also
not quite right either, but it does now actually detect at the conflicts
it was always supposed to.
Panu Matilainen [Thu, 27 Sep 2012 10:11:03 +0000 (13:11 +0300)]
Add doxygen documentation to string pool API
Panu Matilainen [Thu, 27 Sep 2012 10:09:39 +0000 (13:09 +0300)]
Fixup a copy-paste error in rpmdsPool() doxygen markup
Panu Matilainen [Thu, 27 Sep 2012 09:18:39 +0000 (12:18 +0300)]
Fixup string pool prototype argument names to match implementation
- Just a simple s/sidpool/pool/ to match the implementation, "sidpool"
is a leftover from early draft version that somehow made its way
to the master tree.