Re: [Texmacs-dev] Applied patches for TeXmacs-1.0.1.14

[Nix N. Nix]

On Wed, 2003-05-21 at 09:36, Joris van der Hoeven wrote:
> Currently, I think that "release deconstruction" is really not worth
> the effort; looking whether I correctly applied your patches seems
> more than sufficient to me. I rather prefer the construction of
> new releases to the deconstruction of previous ones...

LOL :o) Good point :o)  However, it seems to me that you must /always/,
for /every/ release, tell us which patches you apply - by whatever
channel you choose.  Otherwise, we run the risk of deleting patches
without them ever having been applied and, as with #1400, never applying
patches because you think they had already been applied.

I do think, however, that there is merit to this deconstruction, if
largely academic, and only minorly practical:  The patch history (as I
understand David), after David's work, is such that it maximizes
orthogonality and insulates new dependencies well.

Doing all kinds of diff magic (such as trying to apply a patch, only to
have patch tell us that it thinks it's already applied) I don't think is
a substitute for a definitive statement of applied patches on your part.

I believe a good heuristic should be this (please refine):

Patches on savannah that are still open, that do not say "Joris said:
Applied for x.x.x.xx", and that meet other such criteria (please specify
further) are patches that you should consider for application.

David (and I) will then check to make sure the patch went in OK, and if
not, one of us will post a "patch-patch" to apply the rest of the patch,
and/or correct any "intent errors".