Permalink
Commits on Jun 15, 2018
  1. Merge pull request #271 from ignatenkobrain/python

    mlschroe committed Jun 15, 2018
    bindings: expose repodata_str2dir, repodata_dir2str and repodata_add_…
  2. bindings: expose repodata_str2dir, repodata_dir2str and repodata_add_…

    ignatenkobrain committed Jun 15, 2018
    …dirstr
    
    Signed-off-by: Igor Gnatenko <i.gnatenko.brain@gmail.com>
Commits on Jun 7, 2018
  1. Remove wrong solv_free(data->vincore) in repodata_internalize

    mlschroe committed Jun 7, 2018
    We're extending the old buffer, so the free is always wrong.
Commits on Jun 4, 2018
  1. Merge pull request #262 from j-mracek/const

    mlschroe committed Jun 4, 2018
    Add const for second variable
  2. Merge pull request #266 from ignatenkobrain/typo

    mlschroe committed Jun 4, 2018
    doc: fix typo
Commits on May 15, 2018
  1. Add const for second variable

    j-mracek committed Apr 10, 2018
    The second variable is used as a source for copy and it is not modified
    therefore it should be marked as s const. Then it is more easy to identify what
    is source and what is target and also it is clear that the source variable will
    be not modify by the function.
    
    Additionally it allows to use const map or queue to create a copy or other
    operations.
Commits on May 11, 2018
  1. doc: fix typo

    ignatenkobrain committed May 11, 2018
    Signed-off-by: Igor Gnatenko <i.gnatenko.brain@gmail.com>
Commits on May 7, 2018
  1. Tweak source heuristic in ENABLE_RPMPKG_LIBRPM case

    mlschroe committed May 7, 2018
    Use HEADERGET_MINMEM to make the code faster and save some memory.
  2. Get rid of old cmake cruft

    mlschroe committed May 4, 2018
    The -DVERSION definition is never used, so we can delete it plus
    the cmake_policy setting.
  3. Merge pull request #264 from m-blaha/rpmsourceheuristic

    mlschroe committed May 7, 2018
    source/binary rpm detection heuristic when ENABLE_RPMPKG_LIBRPM used
Commits on May 4, 2018
  1. source/binary rpm detection heuristic when ENABLE_RPMPKG_LIBRPM used

    m-blaha committed May 4, 2018
    The headerIsSource() simply checks a presence of the RPMTAG_SOURCERPM
    tag in package header, which is not really useful in this context.
    This patch provides the same "heuristic" as the librpm uses for
    detecting source/binary rpm.
    
    Relevant bug: https://bugzilla.redhat.com/show_bug.cgi?id=1572338
Commits on Mar 23, 2018
Commits on Mar 1, 2018
Commits on Feb 28, 2018
  1. Merge pull request #257 from ignatenkobrain/seleciton-clone

    mlschroe committed Feb 28, 2018
    bindings: use $self for Selection.clone()
Commits on Feb 27, 2018
  1. Fix saving of branches when refining problems

    mlschroe committed Feb 27, 2018
    This bug caused a mem-leak which led to its discovery...
  2. bindings: use $self for Selection.clone()

    ignatenkobrain committed Feb 26, 2018
    Signed-off-by: Igor Gnatenko <i.gnatenko.brain@gmail.com>
Commits on Feb 26, 2018
  1. Merge pull request #256 from ignatenkobrain/python-rich

    mlschroe committed Feb 26, 2018
    bindings: expose parserpmrichdep()
Commits on Feb 25, 2018
  1. bindings: expose parserpmrichdep()

    ignatenkobrain committed Feb 25, 2018
    Signed-off-by: Igor Gnatenko <i.gnatenko.brain@gmail.com>
Commits on Feb 19, 2018
  1. Work around rpm bug if HEADERIMPORT_FAST is used

    mlschroe committed Feb 19, 2018
    With HEADERIMPORT_FAST, rpm uses the difference between offsets
    to determine the entry length. Unfortunately, the length is
    used as entry count for RPM_BIN_TYPE, so the cound may end up
    to big. So trim it again to match the md5 length.
    
    Thanks to Oleg Girko <ol@infoserver.lv> for investigating this.
Commits on Feb 15, 2018
  1. Fix changes

    mlschroe committed Feb 15, 2018
Commits on Feb 13, 2018
Commits on Feb 12, 2018
  1. Rework fetch phase in fileconflict handling

    mlschroe committed Feb 12, 2018
    We now handle all files of one package in one go instead of
    iterating over every file. In most cases this does not make
    a difference, as the pruned candidate list is very small. But
    it can be big if two packages with the same content get
    installed, e.g. two kernels with same version but different
    rebuild count.
Commits on Feb 8, 2018
  1. fileconflicts: make sure at least one candidate is a new install

    mlschroe committed Feb 8, 2018
    No need to get only installed packages.
Commits on Jan 31, 2018
  1. Remove solver_disableproblem and solver_enableproblem

    mlschroe committed Jan 31, 2018
    They are internal to the solver and can't be used from
    outside. Seems they were added to the libsolv.ver map by
    accident.