Sophie

Sophie

distrib > Mageia > 1 > i586 > by-pkgid > 146b7c9c737452c84aad653bb9781942 > files > 63

ocaml-dose3-devel-2.9.2-2.2457.2.1.mga1.i586.rpm

- modify all test cases to check if the test file exists. If this is 
  the case, run the test, if not ignore the test. (this is to prepare
  a small release tarball with a restricted number of unit tests) 
  Wed Jun 30 2010 Pietro Abate <pietro.abate@pps.jussieu.fr>

- fix ocamlbuild to imply -DHASRPM only when using c_use_rpm in _tags
  and remove -I rpm when building debian binaries / -I deb when building
  rpm binaries
  Thu Jun 17 2010 Pietro Abate <pietro.abate@pps.jussieu.fr>

- fix common/EdosSolver documentation. For some reason only part of it is
  actually generated
  Thu Jun 17 2010 Pietro Abate <pietro.abate@pps.jussieu.fr>

- add minisat backend to the depsolver
** minisat ocaml bindings are now avalaible
  Thu Jun 17 2010 Pietro Abate <pietro.abate@pps.jussieu.fr>

- add picosat backend to the depsolver
** picosat ocaml bindings are now avalaible
  Thu Jun 17 2010 Pietro Abate <pietro.abate@pps.jussieu.fr>

-------------------

- make reason in depsolver lazy. This way we can save a bit of time 
  when building the sat problem

- find a way to add a proper (debian/rpm/real) version to missing 
  dependencies. atm this information is lost when traslating the 
  problem to cudf
** DONE !

- fix the broken dependencies explanation list. The reason list
  sometimes does not include broken packages at the end of a reason
  list. The solver does not provide this information. Actually this is 
  a problem with distcheck.
** DONE !

- when init the depsolver, create a new data structure where all dependencies
  and conflicts are explicit. Then use this data structure when iterate
  over packages. This way we can perform all lookups once and forall.
** DONE

- implement a debdistcheck that builds the sat problem directly from 
  debian Packages without passing through cudf and friends
** WONTFIX