On Tue 2004-12-21 at 22:14:23 -0000, you wrote
[...]
> Personally I accept (or assume?) that core SVN dev's are financially
> supported by an organisation with a heavy unix-bias (nothing wrong
> with that, but I guess win32 is of secondary importance to the
> people that count here. Symbolic links v case-insensitivity anyone?)
Ironic that you should choose the existence of symlinks as an example
for bias of the "core" team.[1] Symlinks were asked for regularly (in
2002 or earlier, see issue 667 [4]) and the dev team didn't implement
it until this year. IIRC, the main reason was because there was no
full concept of what to do on MS Windows and the devs had no time for
finishing that concept themselves, while working on stuff they
considered more important. Sounds familiar?
Symlinks support came to existence mainly through the work of Josh
Pieper this summer, who doesn't belong to the financially supported
dev team you refer to, AFAICT.[2] And if it wasn't for him, there is a
good chance that there still wouldn't be any symlink support.
So, in summary, symlinks happened, because someone who had some
interest and time, tweaked the concept until it withstood the critics[3]
and then did the work. The reason the case issue on MS Windows is not
fixed yet, is because nobody decided it's important enough for him/her
to make it happen. So it has to wait until it crawls up high enough on
the lists of one of the people who has not as much personal interest,
and then finds enough time and energy to take care of it.
That, btw, wasn't meant to say that you shouldn't voice your wishes
here (that's part of the "crawling up on people's list"), but just to
explain that lack of action is no sign of bias against win32,
especially considering that some of the main devs are developing full
time under MS Windows.
Bye,
Benjamin.
[1] Note that the Subversion developers don't like that term very much
when it is used to describe the devs being paid by CollabNet.
There are full committers, which have all the same say. And partial
committers, which have the same say when it's about the stuff they
are committers for.
There are some which are listened to more than others, but that's
only because they usually know more about the topic. So really,
it's about expertise, not about being employed by someone. It just
happens that people spending their whole day on that stuff because
they are paid for it usually build up more expertise than part-time
contributers.
You can easily see that this is really the case when you look at
the dev list at one of the times, where the Collab folks are busy
with non-subversion work. Then life goes on much without them and
suddenly others suggest to cut releases, make the plans, and all
the other work only "authorized" people usually do. That's because
all are authorized. It's a meritocracy.
As someone else recently said on these lists, most of the work
gets done by taking action: make a proposal, build consensus and
then implement. Those not able or willing to do that must rely on
other to do it for them.
[2] While he got full commit in 2003, he really started doing commits
only 3 months before implementing symlinks. True, he did some
aweful work even when he was a casual contributer and still was
submitting his work as patches. Anyhow, his main work began when
he was the co-author of the FSFS backend to ghudson. But
nevertheless, in mid-summer, you could still call him one of the
new-comers, one who had no affiliation with collab.net.
[3] http://svn.haxx.se/dev/archive-2004-06/0986.shtml
[4] http://subversion.tigris.org/issues/show_bug.cgi?id=677
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Dec 22 03:54:02 2004