Robert Widhopf-Fenk <[EMAIL PROTECTED]> writes:

> Sure, but then why should we have the odd numbers? ;-)

The odd numbers should be "a bit more stable than the others".

Even with the new development schema (I try to re-read all patches
before merging them), I committed syntactically incorrect patches.

Odd version numbers should be versions that have been tested for a few
days, and for which no major bug has been found.

> Also I would suggest that the versions used in the archives
> correspond to the major.minor.  And when having one archive
> corresponding to the release/devo version the patchlevel
> should correspond to the patch of the archive.

This is not compatible with my comment above. Here's how I see it:

When releasing the 1.0 version, I'll create a xtla--releases--1.1
branch, which will be a tag-only branch. Each numbered 1.1.x version
will correspond to xtla--releases--1.1--patch-x.

This won't prevent us to have several patches in xtla--main--1.1
between two numbered releases.

(I'm using the future to make it clearer, but this should be a
conditional, of course.)

-- 
Matthieu

Reply via email to