Re: [Savannah-hackers] Mixed up attic for AUCTeX.

[David Kastrup]

Elfyn McBratney <address@hidden> writes:

> Hello David,
> 
> On Wed, Aug 11, 2004 at 11:57:39PM +0200, David Kastrup wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > Sylvain Beucler <address@hidden> writes:
> > 
> > > On 2004.08.11 22:45 David Kastrup wrote:
> > > > Hello,
> > > > when I do a checkout of the CVS archive for AUCTeX without any
> > > > options, I get a current version.
> > > 
> > > Do you mean
> > > cvs -d:ext:address@hidden:/cvsroot/auctex co auctex
> > > ?
> > > 
> > > > However, if I do a checkout with a tag, like -r HEAD or -D 2004/09/01,
> > > > then I get wagonloads of junk from the Attic checked out as well.
> > > 
> > > Can you give us the precise commands, and the files that differ with
> > > the ones you get when using the above command (w/o options)?
> > 
> > Feeling silly today?  Ok, here are the commands:
> 
> Not particularly, you?
> 
> > mkdir junk{a,b}; cd junka
> > cvs -d:ext:address@hidden:/cvsroot/auctex co auctex
> > cd ../junkb
> > cvs -d:ext:address@hidden:/cvsroot/auctex co -r HEAD auctex
> > cd ..
> > 
> > Surprise.
> 
> Indeed, Sylvain was forgetting that we are mind readers and knew exactly
> what commands and/or options you were using.  Bingo.

Well, I simply copied the exact command he gave (which did exactly the
checkout I described), and added the exact option I gave.  Ok, I have
to admit that it is unintuitive that adding "-r HEAD" should make any
difference.  After all, it was probably a side effect from archive
corruption.

> > Here are the results from a diff of the resulting trees:
> >
> > [...]
> > 
> > > > Is it possible to fix this?  It makes it, for example, impossible to
> > > > do an "export" command on the HEAD revision.
> > > 
> > > Can you also check whether you are using a branch in your current
> > > version (CVS/Tag files).
> > 
> > Sigh.  What has the current version to do with it?  Just use the above
> > commands, and you'll have a current version.  There are no CVS/Tag
> > files in it, anyway.
> 
> Sigh.  What doesn't the current version have to do with it?  This could
> have been a bug in a particular client; a bug that was triggered in the
> server due to different mixes of command-line options.

But what has the version I have checked out to do with the versions
for which I specified the command lines?  Or do you mean that I am
using a developer version of CVS itself from a branch?  That was not
at all apparent.

> We didn't know, so we asked you for the version you were using *because*
> you didn't give us enough information.
> 
> > > Last, please specify your CVS version (cvs --version).
> > [...]
> > > Sorry, but we need more details to understand your problem.
> > 
> > Feel free to ask anything else you might require.  But I should think
> > that this problem would be pretty reproducible.
> > [...]
> 
> This is fixed.  Let us know if we missed anything.

If I use the web interface to Savannah at
<URL:http://savannah.gnu.org/cgi-bin/viewcvs/auctex/auctex/#dirlist>,
clicking on the "Attic" link gives me

Python Exception Occurred

Traceback (most recent call last):
  File "/usr/lib/python2.1/viewcvs/viewcvs.py", line 2654, in run_cgi
    main()
  File "/usr/lib/python2.1/viewcvs/viewcvs.py", line 2607, in main
    view_directory(request)
  File "/usr/lib/python2.1/viewcvs/viewcvs.py", line 1155, in view_directory
    fileinfo, alltags = get_logs(full_name, rcs_files, view_tag)
  File "/usr/lib/python2.1/viewcvs/viewcvs.py", line 1102, in get_logs
    raise 'error during rlog: '+hex(status) #+' '+str(chunk)
error during rlog: 0x100


Apart from that, at the first glance the CVS seems to work out now.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum