[Libreoffice] Summary: LibreOffice TSC call - Thur Apr 14th - 16:00 UTC

[Bjoern Michaelsen]

Hi all,

sorry for the late post, here a short summary of the last TSC call:

attendants: Kohei, Petr, Kendy, Norbert, David, Fridrich, Cedric, Bjoern

3.4 status (Petr):
    - beta-1 upload should give us a better picture where we are by user
      feedback
    - beta-2 and further steps should depend on that feedback
    - fdo#33915 is nasty and should be fixed, but we need a
      reproducable scenario (user feedback from beta-1 might help here)

DEV300 imports (Kendy):
    - we will do a very rough trial merge in a feature branch (-> Kendy)
    - we (all) will have a look at that (as a diff to the base) to see:
        - if we can merge that in bulk
        - or selectively
    - decision might differ on a module by module basis

EasyHacks (Bjoern):
    - bugzilla reports still missing on the wiki (-> Bjoern/floeff)
    - we can already make a draft selecting 10-20 tasks now (-> Bjoern)

GSoC (Fridrich/Cedric):
    - seven good applications, some still missing EasyHacks
    - if you mentor a student without an EasyHack, nag him about it

GNU make (Fridrich/Bjoern):
    - on GNU make 3.81 the current var2file implementation triggers
        http://savannah.gnu.org/bugs/?20033
    - GNU make 3.82 rumored to introduce new problems, esp. on cygwin,
      needs investigation
    - using a patched version on release builders deemed acceptable
      (or rather: unavoidable) for now

Unittests:
    - sc ucalc unittest is active again on master and 3.4
    - the rdb setup stuff is still to cumbersome (-> Bjoern)
    - the objectlists stuff breaks on OSX -- naming issue? (-> Bjoern)
    - sw unittest should be activated soon too (-> Bjoern)

split repos vs. one huge repo (Bjoern):
    - push/pull operations should be a lot faster on one repo
    - git blame has no significant performance difference
      (same assumed for git log)
    - stuff that walks the tree might be impacted:
      git status/git commit -A/gitk (-> Bjoern)
    - we would loose the possibility to host a small private repo on
      github, shallow clones considered as an alternative, but might be
      too fragile
    - decision will also need some serious testing on windows

TSC call timeslot:
    - Next call will be 15:00 UTC for real (sorry guys)

Best,

Bjoern

-- 
https://launchpad.net/~bjoern-michaelsen