Alfred Perlstein <[EMAIL PROTECTED]> writes:
> The reason you're talking on the wrong list is that it seems that
> most of maintainers of CVS currently have a 'philosophy'.
> It seems that in the last couple of years philosophy has been the
> thing to strive for rather than program correctness.
>
> Instead of all these warning lights on all the junk that doesn't
> work in CVS it should either be ripped out or the maintainers should
> allow fixes to be incorperated!
>
> Imagine someone's dismay at fixing things like symlink support or
> locks only to be told that "we want it broken".
I should probably stay off this list, as info-cvs has degraded in
recent months; however I will give in to temptation and respond to
what I think is a mischaracterization (albeit unintentional) of the
CVS developers.
Believe me, most CVS developers would love good symlink support, or
permission support, or even unobtrusive locking support. No one ever
said "we want it broken", except perhaps in jest.
The reason a good patch gets ignored is that the developers do not
have time to review the patch with sufficient care. And failure to
review patches has led to problems in the past (mea culpa), so this is
not an imaginary problem -- it is quite possible to do harm by letting
a bad patch in, and distinguishing good from bad is a highly
non-trivial exercise.
> Somehow I feel this to be somewhat of a sadistic thing to do to
> users as well as misusing the whole concept of open source.
Do I have to respond to this? :)
> I can understand a stance where CVS developers refuse to work on
> problem reports with locking or symlinks because they don't want
> to deal with it, but refusing patches that fix real problems is
> pretty bogus.
The stance you say you'd understand is the stance most of us actually
take.
I think you are confusing people who maintain CVS with people who post
a lot to info-cvs. These two groups are distinct, if not entirely
disjoint. In fact, there's usually an inverse correlation -- after
all, anyone who spends all their time posting to this list can't
possibly be doing any real work on CVS, right? :-)
> Seriously folks, the issues with locking and symlinks should up to
> the community to decide... as long as they are the ones doing the
> work and it doesn't impact on non-locking/non-link CVS functionality.
You are free to do the work (and it's a lot of work) and release your
own CVS. If your CVS is sufficiently advanced over the current CVS --
and by "current" I mean "the CVS to which you have been referring in
your mail" -- then I'll be happy to use yours instead.
The code is free and anyone is welcome to maintain a version of it.
There happens to be a group of people, with unfortunately not much
spare time, who collaborate on maintaining the version that can be
downloaded from cyclic.com. There are many changes they'd like to
make to this version, but they don't have time. On the other hand,
they don't want to promiscuously include contributions from all over
the place, because they're afraid the quality of their CVS will suffer
(and they're probably right, considering that most patches to CVS have
bugs and go through several rounds before they're fixed). Hence their
CVS progresses at a slower rate than some users want.
This is not due to territoriality, nor malice, nor an overdeveloped
sense of philosophy or principle. It's simply the predictable outcome
of a situation where people are trying *first* to do no harm, and
*secondly* to improve the program.
> But that doesn't seem possible with the current maintainers being so
> blinded by thier own philosophy. And unfortunatly no one with
> enough resources to actively maintain something has forked a version
> of CVS that is committed to doing the right thing.
Right. You're only mistaken about the motivations of the current
maintainers -- they are limited by time, not vision.
Best,
-Karl
