On Sun, 16 Sep 2007, Malcolm Rowe wrote:
> On Sun, Sep 16, 2007 at 05:48:53PM +0100, Max Bowsher wrote:
> > Is there a reason to allow 0.27.1 but not 0.27.0? (ok, it has a
> > regression, but it doesn't sound like a fatal one).
>
> I don't believe so, no.
Haven't we had reports of various versions of Neon having trouble with
GSSPI auth? IIRC, TortoiseSVN ships with vary specific versions of
Neon and/or compilation options to avoid these problems.
> > For that matter, assuming I volunteered to do the shell/m4 work, is
> > there any known reason not to accept Neon 0.25.anything, 0.26.anything
> > and 0.27.anything ?
>
> Again, no. From what I recall, we have in the past needed to blacklist
> specific patch levels of Neon -- but not for a long time (since I've
> been involved in Subversion). I think the current per-patch whitelist
> is simply a historical leftover that we can remove now.
Sounds like we actually want a blacklist, which can be overridden with
'configure --disable-neon-version-check'.
> > Also, is there any reason for our configure script to continue
> > recommending a specific Neon version when it finds none? I'm not aware
> > of Neon being unstable enough to justify us pointing to the main
> > download page and letting people pick the latest?
> >
>
> Likewise: probably not any more.
>
> Now, Serf, on the other hand... ;-)
Yup, we want a to retaina "recommend versions" list here. ;-)
- application/pgp-signature attachment: stored
Received on Mon Sep 17 20:59:02 2007