On Wed, Sep 08, 1999 at 10:39:42AM -0500, Troy Benjegerdes wrote:
> > Yup, it spanned more than a full screen, which is more than a whole
> > _function_ should be for readability's sake. The 5.3.1 code has a
> > different approach to your solution, but it probably doesn't work, as I
> > haven't tested it on a 2.9.5 compiler.
> 
> I just finished a compile of 5.3.1 with gcc-2.95 and a couple of
> minor patches. I don't know when I'll get a chance to decently test it,
> however.
> 
> If you want the patches, or a log of the build with all warning messages,
> let me know.

Wow, patches are always welcome. I actually just got a machine set up
with gcc-2.95 so I'll see the warnings `live'. We also just upgraded all
our servers to 5.3.1, so the decent testing is already in full effect ;)
(Now I can start delivering my email in Coda again.)

The upgrade went reasonably smooth, all servers came back up fine.

There was at least one case where there was a directory conflict while
the VV's were equal. This used to turn some file in the directory into
an unrepairable conflict. Right now, the client reports 'resource
temporarily unavailable'. Finding the missing server was easy, it
reported RS_LockAndFetch, Get... failed errno 5 (EIO). I shut down the
server, created a new file in the directory to force a conflict, and
restarted the server. Automatic resolution did the rest.

I'll upgrade the testserver either this afternoon or tomorrow, so I
expect some messages about people not being able to connect to the
testserver on the codalist over the next couple of days/weeks.

Jan