This is the mail archive of the cygwin-apps@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Maintainers/Packages List, 2003-11-22


On 2003-11-22T16:09-0500, Christopher Faylor wrote:
) On Sat, Nov 22, 2003 at 05:13:04PM -0000, Daniel Reed wrote:
) >flex                    Christopher Faylor      !!! minor bump (freshmeat lists 2.5.27; we have 2.5.4a)
) See the latest flex release announcement.

It looks like 2.5.27 is the latest stable. The version bump/withdraw from
September was 2.5.31, which is from the "Development" tree according to
freshmeat.

I do not know what might have prompted an incompatibility between patchlevel
releases, but it does look like it's at least partially acknowledged. The
SF.net mailing list archives have messages about incompatibilities
introduced in .31, but do not appear to have any follow-ups to those
messages.


) >initscripts             Sergey Okhapkin         !!! no source and no external-source !!! last updated 2002-11-19
) Shouldn't need source since it is source.

Alrighty; another binary-only.


) something to strive for.  I can imagine some of our more clue challenged
) users latching onto this message and assuming that their problem with
) patchutils is due to the fact that it is now advertised as being out of
) date.

I fully agree, and have been trying to soften the tone of version mismatch
warnings. I intended the warnings to be more advisory than chastising. (I
was actually not fully prepared to send the list out today; I guess I should
have disclaimed that more vigorously.)

On the other hand, speaking as a developer, when I do make a release, there
is usually a decent reason. If nothing else, I believe the maintainers of
[apparently] out-of-date packages should at least be prepared to explain why
the package version is lower than the vendor version. I agree this should
not have to be done repeatedly, though; I marked 2.5.27 as "ignore-before",
so flex will not show up again until/unless a version newer than 2.5.27 is
listed in the "Stable" tree at freshmeat.

I am also going to look into ways to always pull from the stable tree at
freshmeat, instead of just whatever tree the vendor listed first (which is
sometimes the development tree, as is the case with gnugo).

I am not sure what to do with things like Apache, though. Apache has two
"stable" trees, 1.3 and 2.0. Both trees still at least have periodic
maintenance releases, so neither one is necessarily worse than the other,
but eventually the lower of the two *will* stop being updated. It's just
that that day has not yet arrived. Freshmeat lists 2.0 as "Stable" and 1.3
as "1.3", with a separate third "Development" tree.

-- 
Daniel Reed <n@ml.org>	http://naim-users.org/nmlorg/	http://naim.n.ml.org/
There is a lot of food in a supermarket, too, but a supermarket isn't
the best place to hold a dinner party. -- Christopher Faylor


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]