>>>>> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> I just wondered whether it is really such a good idea to have
Andre> all those pre and fix versions around. If we use the
Andre> 'stable/development' distinction we'd could use the even/odd
Andre> scheme as well, if we don't we could use consecutive numbers...
Andre> I personally don't mind having a 1.1.65.
The problem with 'stable/development' is that the two branches diverge
too fast and, not being the linux-kernel, we do not have the manpower
to keep them in sync. You know we just tried that for a year or so,
and after that the development branch just had to be abandonned.
Andre> At least I know that
Andre> known problems from 1.1.47 might be fixed in there whereas the
Andre> relation of 1.1.4fix7 and 1.1.5pre2 is not immediately clear.
I'd be interested to get feedback on whether the current versioning
system is confusing. Currently the scheme is as follows:
- 1.1.4 is a stable release. As soon as it is released, developpers go
back at development work on cvs.
- of course, during the development work, bug are found or reported by
users. If their fix is simple enough (as often, since they are due
to bad coding), it is accumulated in a patch by yours truly. When
enough important fixes have accumulated, I released 1.1.4fix1, which
is a patch against the plain 1.1.4. Other incremental patches have
been released later.
- when enough work has been done on on the development front,
1.1.5pre1 is released, in order to give a glimpse to what 1.1.5 will
be. Normally, after one or two prereleases, the actual 1.1.5
release happens.
Of course, this scheme means that no new feature is added to 1.1.4
in the "fix" series (that's what they are: just boring bug fixes).
So, are there other people who think that this is confusing?
JMarc
PS: the next preversion of LyX 1.1.5, which should have appeared last
friday, will have to wait a bit, since Lars is in vacation...
