Re: [CEDET-devel] Support for sxemacs

[David E. <de...@ra...>]

Nelson Ferreira writes:
> Ok, I changed the inversion regexps a little bit so that the unit tests
> would pass. They were indeed failing.
>
> Now instead of the beta versions being (beta <major#> <minor#> <beta#>)
> it is (beta <major#> <minor#> <beta#> <patch#>)
>
> So (inversion-decode-version  "1.3 beta5") => (beta 1 3 0 5) and
> (inversion-decode-version  "1.3.4") => (full 1 3 4)
> instead of the previous (beta 1 3 5) and (full 1 3 4).
>
> This to me makes a bit more sense, since (full 1 3 4) would seem to be
> an earlier version than a beta for 1.3.5, but surely beta 5 of 1.3.0 is
> much earlier than full 1.3.4.
>
> This kind of comparison was kinda needed because sxemacs has
> major.minor.beta-patch.
>
> My changes do work with all the cases on the unit test, but let me know
> if this change on the tuples is troublesome for you.

I think those changes to the version decoding are fine. Thanks for
looking into that. While not identical, they are at least more in line
with new versioning decoding functions in Emacs proper, so that's a good
thing. I've applied your patches to trunk.

As for getting the grammar compilation to work in (S)XEmacs, have a look
at a discussion in Xemacs.beta from last September. I think Michael
Sperber got pretty far in debugging this; it seems it's some kind of
problem with mode-local running in batch-mode:

http://list-archive.xemacs.org/pipermail/xemacs-beta/2010-September/020048.html

-David