Michael Casadevall <[EMAIL PROTECTED]> tapota :

> 1.  (*) text/plain          ( ) text/html           

This commit will likely create a bunch or merge conflict, as the same
files were updated on the CERN branch.

We need to work differently if you intend to modify heavily files that
were also touched on the CERN branch (one particular thing with
gettext is the fact that new potfiles modify all i18n files). 

I will probably go the rough way for i18n files, meaning that I'll
just overwrite files and not resolve conflicts. For the code, it would
be insane to do it that way; hopefully maybe there wont be any
conflict. But I seen that one your previous changes touched the
trackers, and most of the changes on the branch happens on the
trackers. I hope we will not have a bad surprise the merge day (since 
unfortunately, it wont be possible  to test the tracker with
your commit without prematurely merging into the trunk).

More generally, when a branch is open and active on a specific part of
the code, the best is probably to avoid touching that part of the code
or asking to the persons that work on the branch if it possible to
commit directly on that branch (or even to branch on the branch).

Sorry if this mail sound archs, I'm just mentioning frankly potential
issues of this commit. I'm not attacking Michael work. If he made that
commit on the trunk, it is not his fault but rather mine, who forgot
to explicitely say that commits on the trunk of files also touched on
the CERN branch could be a pain the merge day. I realize I made this
mistake and that's why I reply on savane-dev about this commits, not
to patronize Michael but just to do what I should have done before:
raising the problems of concurrent commits.

There's nothing to do about what was done before and I do not think it
will be a big deal (even if there are conflicts, we are only talking
about one code commit, should be fast to fix) but I'd like to avoid
this thing getting really problematic.

Sum up: the CERN branch mostly affect frontend trackers code, but
affect most general the whole frontend. Commits on the frontend should
be avoided. Commiting on the CERN branch is way preferred (it wont be
the case after November the 21th) so it can be directly tested and
wont create merge conflicts. If your commit is about the contend of 
include/trackers*, send me a mail so I'll tell you the exact status of
the file (currently being modified or else).

The general rules is that bugfix commits are acceptable on the
trunk. But since it is the CERN branch that is tested and was directly
derived from the trunk, even bugfix and cosmetics are best being
commit on the CERN branch (which could be consired as 1.0.5-dev
branch).

I hope this clarification wont discourage Michael about commits ;)
I perfectly understand how painy it could be to have plenty of "rules"
to follow just for a few commits; but the end of it not playing
generals and soldiers, only working more efficiently.

Regards,


-- 
Mathieu Roy

  +---------------------------------------------------------------------+
  | General Homepage:           http://yeupou.coleumes.org/             |
  | Computing Homepage:         http://alberich.coleumes.org/           |
  | Not a native english speaker:                                       |
  |     http://stock.coleumes.org/doc.php?i=/misc-files/flawed-english  |
  +---------------------------------------------------------------------+

Reply via email to