apt.git
8 years agoAdd bunch of missing includes to fix build with gcc 4.7 master
Panu Matilainen [Wed, 26 Sep 2012 18:31:36 +0000 (21:31 +0300)]
Add bunch of missing includes to fix build with gcc 4.7

9 years agoRemove redundant IsDatabase() variants
Panu Matilainen [Sat, 16 Apr 2011 18:35:06 +0000 (21:35 +0300)]
Remove redundant IsDatabase() variants
- Only RPMDBHandler needs to override it, everything else can
  use default from base class

9 years agoRemove swig-generated content + ancient patch to swig
Panu Matilainen [Wed, 13 Apr 2011 19:30:05 +0000 (22:30 +0300)]
Remove swig-generated content + ancient patch to swig
- Apt used to require a local hack to swig, which is why pre-generated
  content has been needed but that shouldn't be the case anymore.
  Knock wood.

9 years agoMassage swig interface a bit to get this to build on gcc 4.6.0 / swig 2.0.2
Panu Matilainen [Wed, 13 Apr 2011 19:28:12 +0000 (22:28 +0300)]
Massage swig interface a bit to get this to build on gcc 4.6.0 / swig 2.0.2

9 years agoAnd finally, adding SHA256 support becomes a one-liner patch
Panu Matilainen [Wed, 13 Apr 2011 17:58:23 +0000 (20:58 +0300)]
And finally, adding SHA256 support becomes a one-liner patch
- This hardwired list is still dumb of course since we know perfectly
  well which hashes we'd like from the acquirer methods but currently
  there's no means to communicate that to the methods.

9 years agoGeneralize Hashes helper class to support arbitrary number of hashes
Panu Matilainen [Wed, 13 Apr 2011 17:57:08 +0000 (20:57 +0300)]
Generalize Hashes helper class to support arbitrary number of hashes

9 years agoAdd various missing bits to raptHash class
Panu Matilainen [Tue, 12 Apr 2011 16:17:21 +0000 (19:17 +0300)]
Add various missing bits to raptHash class
- A destructor was missing (but not strictly necessary assuming
  Result() is always called)
- Add assignment operator + copy-constructor
- Do lazy init of the digest type at time of Add() where we can
  actually return the error (without introducing exceptions)
- Remember the hash type string + add getter method for it

9 years agoRemove now unnecessary raptHash constructor variant
Panu Matilainen [Tue, 12 Apr 2011 15:45:25 +0000 (18:45 +0300)]
Remove now unnecessary raptHash constructor variant

9 years agoAvoid leaking rpm enums into apt space
Panu Matilainen [Tue, 12 Apr 2011 15:44:33 +0000 (18:44 +0300)]
Avoid leaking rpm enums into apt space

9 years agoReplace a couple of leftover md5/sha1 summation uses
Panu Matilainen [Tue, 12 Apr 2011 15:44:17 +0000 (18:44 +0300)]
Replace a couple of leftover md5/sha1 summation uses

9 years agoOops, forgotten repomd -> apt hash name mapping
Panu Matilainen [Tue, 12 Apr 2011 14:53:06 +0000 (17:53 +0300)]
Oops, forgotten repomd -> apt hash name mapping

9 years agoRemember repo file hash types in the index checksum struct
Panu Matilainen [Tue, 12 Apr 2011 12:49:47 +0000 (15:49 +0300)]
Remember repo file hash types in the index checksum struct
- This generalizes things a bit and lets us get rid of silly
  GetCheckMethod() calls here and there

9 years agoTeach pkgAcqFile about different hash types, use for sources
Panu Matilainen [Tue, 12 Apr 2011 12:31:30 +0000 (15:31 +0300)]
Teach pkgAcqFile about different hash types, use for sources
- For arbitrarily downloaded url's we can't verify any checksums
  but sources do come from repositories and have checksum + type
  available, use it.

9 years agoAcquire-item hashification, step 2
Panu Matilainen [Tue, 12 Apr 2011 12:25:33 +0000 (15:25 +0300)]
Acquire-item hashification, step 2
- All acquirers might need to deal with checksum verification, make
  the hash + hashtype members into base pkgAcquire::Item() class,
  call them checksums/hashes and adjust callers'n stuff as needed.

9 years agoAcquire-item hashification, step 1
Panu Matilainen [Tue, 12 Apr 2011 11:27:23 +0000 (14:27 +0300)]
Acquire-item hashification, step 1
- Rename the acquire-item Done() checksum arguments from
  md5-variants -> AcqHash (as that's the checksum the acquirer got,
  what we expect is stored in the item)

9 years agoAdd methods for dealing with varying hash type
Panu Matilainen [Tue, 12 Apr 2011 09:06:00 +0000 (12:06 +0300)]
Add methods for dealing with varying hash type
- Repomd repositories carry checksum type per package record,
  and while in practise it should be the same for all packages we dont
  really know...
- Also unify the repomd hash string -> apt hash string conversion to
  one helper function

9 years agoReplace algorithm-specific md5 and sha1 methods with a common Hash() method
Panu Matilainen [Tue, 12 Apr 2011 08:29:10 +0000 (11:29 +0300)]
Replace algorithm-specific md5 and sha1 methods with a common Hash() method
- The current situation was just broken anyway as we were simply
  returning SHA1 when MD5 was requested for repomd repositories, and
  archives only really have one checksum type for individual files,
  so its not as we could actually choose what to use.

9 years agoUse rhash directly in the hashes wrapper class
Panu Matilainen [Tue, 12 Apr 2011 07:17:35 +0000 (10:17 +0300)]
Use rhash directly in the hashes wrapper class
- Except for test/hash.cc, MD5Summation and SHA1Summation are now
  unused and could be removed

9 years agoRemove no longer necessary rpm compat goo
Panu Matilainen [Fri, 8 Apr 2011 06:59:53 +0000 (09:59 +0300)]
Remove no longer necessary rpm compat goo

9 years agoConvert genpkglist and gensrclist to new rpm interfaces
Panu Matilainen [Fri, 8 Apr 2011 06:25:52 +0000 (09:25 +0300)]
Convert genpkglist and gensrclist to new rpm interfaces
- Wooo, apt finally compiles without -D_RPM_4_4_COMPAT crap

9 years agoOops, python Makefile went missing
Panu Matilainen [Thu, 7 Apr 2011 17:17:31 +0000 (20:17 +0300)]
Oops, python Makefile went missing
- That particular Makefile is not autofoo-integrated so it went
  missing in commit 4f4ebc2c8f8bf697110bbb80a6adf4fe0f84fb64. Oops,
  resurrect for now.

9 years agoRemove updateinfo remnants, broken since apt-rpm 0.3.x days
Panu Matilainen [Thu, 7 Apr 2011 16:08:38 +0000 (19:08 +0300)]
Remove updateinfo remnants, broken since apt-rpm 0.3.x days

9 years agoChainsaw support for rpm < 4.6.0 out for real
Panu Matilainen [Thu, 7 Apr 2011 15:25:48 +0000 (18:25 +0300)]
Chainsaw support for rpm < 4.6.0 out for real

9 years agoRequire rpm version that supports pkg-config
Panu Matilainen [Thu, 7 Apr 2011 14:59:05 +0000 (17:59 +0300)]
Require rpm version that supports pkg-config
- In terms of upstream releases, this means we now require rpm >= 4.6.0.
  Sorry folks, what little time I have available for apt, I dont want
  to spend wrestling with prehistoric rpm versions.

9 years agoRemove remaining rpm5-specific config foo
Panu Matilainen [Thu, 7 Apr 2011 12:30:18 +0000 (15:30 +0300)]
Remove remaining rpm5-specific config foo

9 years agoRip out broken rpm runtime dependency foobar
Panu Matilainen [Thu, 7 Apr 2011 12:18:19 +0000 (15:18 +0300)]
Rip out broken rpm runtime dependency foobar
- This stuff is only half-implemented, mostly broken and only exists
  in rpm5 which we dont work with anyway etc

9 years agoRip out repackage support
Panu Matilainen [Thu, 7 Apr 2011 12:14:49 +0000 (15:14 +0300)]
Rip out repackage support
- Current rpm doesn't support it at all, we dont need to care either

9 years agoRip out selinux "support"
Panu Matilainen [Thu, 7 Apr 2011 12:10:03 +0000 (15:10 +0300)]
Rip out selinux "support"
- Its up to rpm to setup the file contexts as necessary, not our
  headache. I frankly dont care about rpm <= 4.4.x at this point...

9 years agoMeh, leftover autogenerated foo
Panu Matilainen [Thu, 7 Apr 2011 12:07:33 +0000 (15:07 +0300)]
Meh, leftover autogenerated foo

9 years agoRemove the bundled copy of lua from repository too
Panu Matilainen [Thu, 7 Apr 2011 11:46:41 +0000 (14:46 +0300)]
Remove the bundled copy of lua from repository too
- This was supposed to be gone since ages ago... but apparently not

9 years agoRemove AM_MAINTAINER_MODE, me no like
Panu Matilainen [Thu, 7 Apr 2011 11:44:35 +0000 (14:44 +0300)]
Remove AM_MAINTAINER_MODE, me no like

9 years agoRemove autogenerated autofoo
Panu Matilainen [Thu, 7 Apr 2011 11:36:18 +0000 (14:36 +0300)]
Remove autogenerated autofoo

9 years agoBump version to 0.5.15lorg3.95
Panu Matilainen [Thu, 7 Apr 2011 10:10:47 +0000 (13:10 +0300)]
Bump version to 0.5.15lorg3.95

9 years agoRemember to update POTFILES.in wrt hash changes
Panu Matilainen [Thu, 7 Apr 2011 10:05:12 +0000 (13:05 +0300)]
Remember to update POTFILES.in wrt hash changes

9 years agoAdjust to newer git
Panu Matilainen [Thu, 7 Apr 2011 10:02:39 +0000 (13:02 +0300)]
Adjust to newer git

9 years agoAmbiguous || && if
Panu Matilainen [Wed, 6 Apr 2011 18:11:07 +0000 (21:11 +0300)]
Ambiguous || && if

9 years agoUnused/leftover variable
Panu Matilainen [Wed, 6 Apr 2011 18:08:53 +0000 (21:08 +0300)]
Unused/leftover variable

9 years agoActually use the calculated Partial value in shell update
Panu Matilainen [Wed, 6 Apr 2011 18:08:13 +0000 (21:08 +0300)]
Actually use the calculated Partial value in shell update

9 years agoShut up a few complaints about if-ambiguosity
Panu Matilainen [Wed, 6 Apr 2011 18:04:17 +0000 (21:04 +0300)]
Shut up a few complaints about if-ambiguosity

9 years agoFix up callback bogosity
Panu Matilainen [Wed, 6 Apr 2011 17:58:17 +0000 (20:58 +0300)]
Fix up callback bogosity

9 years agoFix initialization order to shut up g++
Panu Matilainen [Wed, 6 Apr 2011 17:55:47 +0000 (20:55 +0300)]
Fix initialization order to shut up g++

9 years agoLeftover variable
Panu Matilainen [Wed, 6 Apr 2011 17:53:47 +0000 (20:53 +0300)]
Leftover variable

9 years agoWe dont have any "binaries" tags in rpm, rip the bitrotten goo
Panu Matilainen [Wed, 6 Apr 2011 17:08:56 +0000 (20:08 +0300)]
We dont have any "binaries" tags in rpm, rip the bitrotten goo

9 years agoEliminate unnecessary strdup()'ing from rpm version parsing
Panu Matilainen [Wed, 6 Apr 2011 15:53:40 +0000 (18:53 +0300)]
Eliminate unnecessary strdup()'ing from rpm version parsing
- Make a copy of the string to stack in DoCmpVersion() already and let
  ParseVersion() modify that, returning const pointers to our local
  data instead of strdup()'ing and freeing many many tiny bits.
- Additionally remove ParseVersion() from rpmVersioningSystem class,
  this is just a local, static helper.

9 years agoClean up sqlite wrapper a bit
Panu Matilainen [Wed, 6 Apr 2011 12:29:39 +0000 (15:29 +0300)]
Clean up sqlite wrapper a bit
- Remove leftovers, simplify
- Grab column names right after preparing the sql statement instead of
  the step() busypath

9 years agoSimplify the sqlite query interface
Panu Matilainen [Wed, 6 Apr 2011 11:20:38 +0000 (14:20 +0300)]
Simplify the sqlite query interface
- We've no need to differentiate query creation + preparation,
  pass the sql to Query() as argument already and hide the details
  in the sqlite wrapper (it's still a bit stupid, it shouldn't be
  possible to instanciate SqliteQuery objects directly but for now...)

9 years agoUse sqlite binding for dependency, filelist and changelog queries
Panu Matilainen [Wed, 6 Apr 2011 10:50:36 +0000 (13:50 +0300)]
Use sqlite binding for dependency, filelist and changelog queries
- Gives a nice fat speed-boost for sqlite operations by avoiding
  unnecessary re-re-re-re-evaluating of the sql statememts when
  only pkgKey changes.

9 years agoAdd integer binding support to sqlite helper
Panu Matilainen [Wed, 6 Apr 2011 10:41:59 +0000 (13:41 +0300)]
Add integer binding support to sqlite helper

9 years agoAvoid unnecessary extra variable, saner sql for counting pkgs
Panu Matilainen [Wed, 6 Apr 2011 08:59:58 +0000 (11:59 +0300)]
Avoid unnecessary extra variable, saner sql for counting pkgs
- Not sure how much of a difference if any it makes, but we only need
  to count of pkgIds, not everything. Also we already have a tmp
  variable for SqliteQuery, just reuse it for the pkg count.

9 years agoClean up leftover junk
Panu Matilainen [Tue, 5 Apr 2011 13:02:01 +0000 (16:02 +0300)]
Clean up leftover junk

9 years agoPermit arbitrary checksum types in VerifyChecksums()
Panu Matilainen [Tue, 5 Apr 2011 11:54:17 +0000 (14:54 +0300)]
Permit arbitrary checksum types in VerifyChecksums()
- just pass the checkmethod string to raptHash() constructor and let
  it deal with the details

9 years agoAdd alternative raptHash() constructor taking string name of the hash
Panu Matilainen [Tue, 5 Apr 2011 11:51:07 +0000 (14:51 +0300)]
Add alternative raptHash() constructor taking string name of the hash

9 years agoTeach repomd bits about sha256
Panu Matilainen [Tue, 5 Apr 2011 11:48:57 +0000 (14:48 +0300)]
Teach repomd bits about sha256
- The repomd always carried its checksum types but we've been
  mostly lazy and just hardwired sha1 there.

9 years agoRemember hash result as Result() might be called more than once
Panu Matilainen [Tue, 5 Apr 2011 11:06:44 +0000 (14:06 +0300)]
Remember hash result as Result() might be called more than once
- also plug memory leak, rpmDigestFinal() returns malloced data

9 years agoEliminate redundant and buggy offset tracking in sqlite query
Panu Matilainen [Tue, 5 Apr 2011 10:01:40 +0000 (13:01 +0300)]
Eliminate redundant and buggy offset tracking in sqlite query
- The handler has its own offset tracking, sqlite doesn't need
  to care about it: it only provides Step() and Rewind() operations.
  This also fixes world-breaking off-by-one.

9 years agoReplace local sha1 and md5 implementations with rpm's hash functionality
Panu Matilainen [Tue, 5 Apr 2011 08:43:40 +0000 (11:43 +0300)]
Replace local sha1 and md5 implementations with rpm's hash functionality
- implement an algorithm agnostic c++ interface to rpm's digest functions,
  based on the former md5/sha1 summation class interface
- minimally "port" md5 + sha1 summation classes on top of the new class
- the result is always used in string format, not bothering to wrap
  MD5/SHA1SumValue class fully, just typedef them as strings and fix
  users

9 years agoEliminate unused MD5/SHA1SumValue methods
Panu Matilainen [Tue, 5 Apr 2011 08:37:21 +0000 (11:37 +0300)]
Eliminate unused MD5/SHA1SumValue methods

9 years agoUse unordered_set (from tr1) for SeenPackages if available
Panu Matilainen [Tue, 5 Apr 2011 08:31:54 +0000 (11:31 +0300)]
Use unordered_set (from tr1) for SeenPackages if available

9 years agoUse STL unordered_map if available instead of nonstd hash_map
Panu Matilainen [Tue, 5 Apr 2011 08:10:12 +0000 (11:10 +0300)]
Use STL unordered_map if available instead of nonstd hash_map

9 years agoA set is more appropriate than hash/map for SeenPackages
Panu Matilainen [Tue, 5 Apr 2011 07:35:30 +0000 (10:35 +0300)]
A set is more appropriate than hash/map for SeenPackages
- We dont have any meaningful associated data here, the hash/map
  has only been used as a set anyway. We'll want an unordered set
  here where supported but for now...
- Also use STL strings instead of C strings

9 years agoUse STL strings for arch score typemaps as well
Panu Matilainen [Tue, 5 Apr 2011 06:53:09 +0000 (09:53 +0300)]
Use STL strings for arch score typemaps as well

9 years agoAdd missing <algorithm> include, masked by hash_map usage
Panu Matilainen [Tue, 5 Apr 2011 06:47:09 +0000 (09:47 +0300)]
Add missing <algorithm> include, masked by hash_map usage

9 years agoUse STL strings for arch score methods
Panu Matilainen [Tue, 5 Apr 2011 06:32:14 +0000 (09:32 +0300)]
Use STL strings for arch score methods

9 years agoForce read-only access to sqlite
Panu Matilainen [Tue, 29 Mar 2011 18:16:46 +0000 (21:16 +0300)]
Force read-only access to sqlite
- We never want to create any db's (at least currently), so if
  a file is not there we want an error, not an empty db created.
  Using the newer sqlite3_open_v2() interface allows specifying this.

9 years agoUse saner sqlite access model
Panu Matilainen [Tue, 29 Mar 2011 17:38:47 +0000 (20:38 +0300)]
Use saner sqlite access model
- The sqlite repo handler wants the data row by row, and sqlite's
  core prepare -> step -> finalize API fits our needs like a glove.
  Instead of allocating ridiculous amounts of memory and copying
  stuff around even more, use this step-by-step access model instead.
- Clean up related cruft a bit while at it - eg. query's Size() is not
  meaningful with the stepping API, use sql count() to get the
  number of packages in the handler instead.

10 years agoRip out support for RPM v3. Time to let go of something...
Panu Matilainen [Tue, 23 Nov 2010 19:45:28 +0000 (21:45 +0200)]
Rip out support for RPM v3. Time to let go of something...

10 years agoMake the package hash table MUCH larger
Panu Matilainen [Mon, 1 Nov 2010 14:23:26 +0000 (16:23 +0200)]
Make the package hash table MUCH larger
- With Fedora 14 repositories, the largest hash chains could top 30, with
  average lookups over five. No wonder it was slow. The hash table
  size + function could probably be further tuned but just having
  a larger table helps immensely.

10 years agoPass pkgCache::FindPkg() name by const reference instead of copy
Panu Matilainen [Mon, 1 Nov 2010 12:59:19 +0000 (14:59 +0200)]
Pass pkgCache::FindPkg() name by const reference instead of copy
- The real performance issues are elsewhere (hash table exploding)
  but making string copies on a busy path doesn't help.

10 years agoSlightly less WTF'ery sqlite filelist generation
Panu Matilainen [Sun, 31 Oct 2010 19:35:20 +0000 (21:35 +0200)]
Slightly less WTF'ery sqlite filelist generation
- Dunno what I've been thinking (drinking?) when writing the original
  version (possibly "hey it seems to work")... but way way way too much
  string searching for the simple operation of finding "/" boundaries.
  Speeds up filelist processing by a considerable margin...

10 years agoOptimize rpmlib() dependency lookups
Panu Matilainen [Sun, 31 Oct 2010 15:46:07 +0000 (17:46 +0200)]
Optimize rpmlib() dependency lookups
- rpmdsRpmlib() is a fairly expensive call, only call it once and
  keep the result in a static variable.

10 years agoFix build with rpm.org HEAD
Panu Matilainen [Sat, 23 Oct 2010 08:38:42 +0000 (11:38 +0300)]
Fix build with rpm.org HEAD

10 years agoUse const iterators throughout apt-pkg/rpm/ where possible
Panu Matilainen [Mon, 11 Oct 2010 09:58:13 +0000 (12:58 +0300)]
Use const iterators throughout apt-pkg/rpm/ where possible

10 years agoOptimize sqlite handler filelist generation a bit
Panu Matilainen [Mon, 11 Oct 2010 09:05:44 +0000 (12:05 +0300)]
Optimize sqlite handler filelist generation a bit
- move string instantiation out of the loop, call
- avoid creating temporaries unnecessarily

10 years agoOptimize sqlite handler PRCO() a bit
Panu Matilainen [Mon, 11 Oct 2010 08:53:54 +0000 (11:53 +0300)]
Optimize sqlite handler PRCO() a bit
- move string instantiation out of the loop, call .clear() as necessary
- avoid creation temporaries

10 years agoOptimize rpm handler EVR calculation a bit
Panu Matilainen [Mon, 11 Oct 2010 08:25:22 +0000 (11:25 +0300)]
Optimize rpm handler EVR calculation a bit
- Avoid generating temporaries 'n stuff

10 years agoSwitch a few callers to the return via reference variants of sqlite getters
Panu Matilainen [Mon, 11 Oct 2010 08:21:18 +0000 (11:21 +0300)]
Switch a few callers to the return via reference variants of sqlite getters

10 years agoAdd new getter variants to sqlite, returning the data via reference args
Panu Matilainen [Mon, 11 Oct 2010 08:02:41 +0000 (11:02 +0300)]
Add new getter variants to sqlite, returning the data via reference args

10 years agoBump cache limit up to 256MB, should be enough for a while
Panu Matilainen [Mon, 11 Oct 2010 07:17:12 +0000 (10:17 +0300)]
Bump cache limit up to 256MB, should be enough for a while

10 years agoBe selective about what we select from repomd databases
Panu Matilainen [Sun, 10 Oct 2010 20:45:47 +0000 (23:45 +0300)]
Be selective about what we select from repomd databases
- Only select what we actually use instead of just grabbing everything.
  The amount of data can be enormous, so having less data to shuffle
  around speeds things up and wastes less memory.

10 years agoAnd yet more pass-by-const-reference fixes, now in cache generation
Panu Matilainen [Sun, 10 Oct 2010 19:47:30 +0000 (22:47 +0300)]
And yet more pass-by-const-reference fixes, now in cache generation

10 years agoYet more pass-by-reference fixes: xml parsing utilities
Panu Matilainen [Sun, 10 Oct 2010 19:00:27 +0000 (22:00 +0300)]
Yet more pass-by-reference fixes: xml parsing utilities

10 years agoRemove couple of leftover method prototypes
Panu Matilainen [Sun, 10 Oct 2010 18:24:16 +0000 (21:24 +0300)]
Remove couple of leftover method prototypes

10 years agoMore pass-by-reference tweaks to package data stuff
Panu Matilainen [Sun, 10 Oct 2010 18:12:33 +0000 (21:12 +0300)]
More pass-by-reference tweaks to package data stuff

10 years agoEliminate unused LessPred comparison struct
Panu Matilainen [Sun, 10 Oct 2010 17:51:29 +0000 (20:51 +0300)]
Eliminate unused LessPred comparison struct

10 years agoPass to-be-hashed strings by const reference to avoid copying, woops
Panu Matilainen [Sun, 10 Oct 2010 17:44:36 +0000 (20:44 +0300)]
Pass to-be-hashed strings by const reference to avoid copying, woops

10 years agoReplace global index size cache with per-index mutable private member
Panu Matilainen [Sun, 10 Oct 2010 16:49:40 +0000 (19:49 +0300)]
Replace global index size cache with per-index mutable private member

10 years agoMake rpm handler internals (explicitly) private where possible
Panu Matilainen [Sun, 10 Oct 2010 10:17:10 +0000 (13:17 +0300)]
Make rpm handler internals (explicitly) private where possible

10 years agoAdd and use string-aware AddCRC16() variant
Panu Matilainen [Sat, 9 Oct 2010 06:52:10 +0000 (09:52 +0300)]
Add and use string-aware AddCRC16() variant

10 years agoPass strings by (const) reference to sqlite getters
Panu Matilainen [Sat, 9 Oct 2010 06:06:18 +0000 (09:06 +0300)]
Pass strings by (const) reference to sqlite getters

10 years agoPass strings to hash by const reference to avoid copying
Panu Matilainen [Sat, 9 Oct 2010 05:59:14 +0000 (08:59 +0300)]
Pass strings to hash by const reference to avoid copying

10 years agoConstify raptHeader getter methods
Panu Matilainen [Fri, 8 Oct 2010 18:42:06 +0000 (21:42 +0300)]
Constify raptHeader getter methods

10 years agoEliminate uses of "using namespace std" from headers
Panu Matilainen [Fri, 8 Oct 2010 18:38:58 +0000 (21:38 +0300)]
Eliminate uses of "using namespace std" from headers

10 years agoMass constify rpmhandler getter methods
Panu Matilainen [Fri, 8 Oct 2010 17:23:46 +0000 (20:23 +0300)]
Mass constify rpmhandler getter methods

10 years agoRpm initialization sanity
Panu Matilainen [Sat, 2 Oct 2010 09:13:58 +0000 (12:13 +0300)]
Rpm initialization sanity
- Initialize rpm configuration at rpmSystem constructor, free
  on destructor. Seems pretty obvious, duh.

10 years agoAccept both "HTTP/1.0 304 Not Modified" and "HTTP/1.0 304"
Sergey [Fri, 1 Oct 2010 17:53:35 +0000 (20:53 +0300)]
Accept both "HTTP/1.0 304 Not Modified" and "HTTP/1.0 304"

10 years agoFix off-by-one in sqlite API wrapper (RhBug:428855)
Sergey [Fri, 1 Oct 2010 17:50:17 +0000 (20:50 +0300)]
Fix off-by-one in sqlite API wrapper (RhBug:428855)

10 years agoCorrectly print a replacement line (RhBug:428857)
Sergey [Fri, 1 Oct 2010 17:48:36 +0000 (20:48 +0300)]
Correctly print a replacement line (RhBug:428857)

Attached pacth "fixes" the problem. Now output should look like:
The following packages will be REPLACED:
   compat-gcc (7.3-2.96.128) (by compat-libgcc-296)
   sysklogd (1.4.2-5.fc7) (by rsyslog)

It does not print versions of replacement packages here,
but you can see them in 'NEW packages' section.

10 years agoUpdate autofoo
Panu Matilainen [Fri, 1 Oct 2010 17:04:09 +0000 (20:04 +0300)]
Update autofoo

10 years agoMacro kludgery to get it to build with rpm.org HEAD
Panu Matilainen [Fri, 1 Oct 2010 17:03:38 +0000 (20:03 +0300)]
Macro kludgery to get it to build with rpm.org HEAD

10 years agoheaderUnlink() is wrong, use headerFree() instead
Panu Matilainen [Fri, 1 Oct 2010 16:35:20 +0000 (19:35 +0300)]
headerUnlink() is wrong, use headerFree() instead
- headerUnlink() never actually frees the data, only decrements
  refcount.