[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: SVN install on Fedora Core?

From: Robert P. J. Day <rpjday_at_mindspring.com>
Date: 2004-03-23 13:44:31 CET

On Tue, 23 Mar 2004, Joe Cooper wrote:

> If you're using bleeding edge software (which Subversion still is, due
> to its bleeding edge dependencies and despite the 1.0 version number),
> you're going to bleed. Can't be much clearer than that. ;-)

nope, sorry, that don't wash. if you go to the trouble of creating a
repository, making it publicly available, and specifically naming some of
those rpms with the suffix of "fc1.i386.rpm" suggesting that they're FC1
compatible, then you have a real ethical obligation to make sure they
*are* FC1-compatible, and that doing an update from that repo won't bork
your system (which it does). obviously, you and i have different
philosophies of the responsibilities of repository maintainers, and it's
clear we're not going to agree here. so be it.
> BTW-David's repository really isn't /that/ incompatible.

either a repo is compatible or it isn't. saying it's only a little bit
incompatible is like saying you're only a little bit pregnant. or saying
that it made my system only a little bit broken.

> I've had only a few minor issues along the way--and those only because I
> use a custom Apache and we rely on lots of modules (ssl, python, etc.).
> You're making a lot out of problems that don't seem to be affecting very
> many folks.

as i've pointed out, this will affect anyone running an FC1 system. i'm
pretty sure that counts as "many folks".

> And while we're on the subject...the FC2 Test 1 directory
> will /also/ break your system in some circumstances. I'm sure you're
> aware of this, since I recall seeing posts from you on the Fedora
> Testing list, and there has been much discussion of various breakages
> due to Xorg vs. XFree86, etc. lately.

there is a big difference here. while some of the rpms in the FC
rawhide/development stream will undoubtedly break your system, they're not
*supposed* to and, if they do, they'll be fixed. that's what the testing
is all about -- identifying those RPMs that don't work and repairing them.

this is different from the summersoft repo, which is guaranteed to break
your FC1 system *by* *design* since it incorporates FC1-incompatible rpms
as a matter of course. that updating from the summersoft repo will mess
up your FC1 system is not an accident -- it's not just an "oopsie", it's a
guarantee. see the difference?
> BTW2-I don't see anything on David's repositories claiming to be
> designed to work cleanly with FC1 systems.

come now, that's taking splitting hairs to a ridiculous extreme. it's a
publicly-available repository, with RPM files, that's set up as a yum
repository, where some of the RPMs even incorporate the phrase "fc1" in
their name. under the circumstances, should it be unreasonable for
someone to assume that it represents an FC1-compatible yum repo from
which i can update?


p.s. at this point, i think we've flogged this thread pretty much to
death and we can agree to disagree. if anything came of it, i guess the
only point i want to get across is that, at the moment, anyone who runs an
FC1 system and updates against the summersoft repo does so at their own
risk. time to move on.

To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue Mar 23 13:47:21 2004

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.