Re: Fix in vc-ignore-dir-regexp

[Stefan Monnier]

> I had reported a related bug many months ago about VC being too
> aggressive with defaulting to the number of versioning systems
> that it wants to support.

> I suggest that we take out all except svn,, cvs and rcs from that
> list.

> In particular vc systems that need to ascend the directory tree
> looking for vc directories should not appear in the default list.

Given that all recent vc systems (other than Subversion) need to ascend the
directory tree, I think it's too drastic.

But it did occur to me recently that maybe a good heuristic to decide when
to stop looking up is to check the directory's owner (and stop when it
changes).

In the mean time I've added /.../ to the vc-ignored-dir regexp as suggested.
I didn't add the /gsa/ one because I've never heard of the gsa filesystem
and Google didn't seem to have heard of it either.


        Stefan


>>>>> "Richard" == Richard Stallman <address@hidden> writes:
Richard> Would an expert in VC please install this if it is
Richard> correct?  ------- Start of forwarded message -------
Richard> X-Spam-Status: No, score=0.5 required=5.0
Richard> tests=DNS_FROM_RFC_ABUSE, UNPARSEABLE_RELAY
Richard> autolearn=no version=3.1.0 From: Smith
Richard> <address@hidden> MIME-Version: 1.0
Richard> Content-Type: text/plain; charset=us-ascii Date:
Richard> Wed, 25 Apr 2007 18:18:52 -0500 To:
Richard> address@hidden Subject: stat problems
Richard> Reply-To: address@hidden
Richard> 
Richard> I am running emacs 22.0.95 on AIX 5.3 (the latest
Richard> levels).
Richard> 
Richard> The special feature is IBM has dfs.  I guess it is
Richard> normal but the dfs file system at IBM starts out
Richard> with /.../path/to/files
Richard> 
Richard> When I open a file, emacs hangs for a long time.  I
Richard> put in some fprintf's into file-exists-p (dbx told
Richard> me that is where I was hanging).  The output looks
Richard> like this:
Richard> 
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/aix52Q/{arch}/=tagging-method'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/{arch}/=tagging-method'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/{arch}/=tagging-method'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/{arch}/=tagging-method' stat
Richard> for '/.../austin.ibm.com/{arch}/=tagging-method'
Richard> stat for '/.../{arch}/=tagging-method' stat for
Richard> '/{arch}/=tagging-method' ...snip...  stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/aix52Q/MCVS/CVS'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/MCVS/CVS' stat
Richard> for '/.../austin.ibm.com/fs/projects/MCVS/CVS' stat
Richard> for '/.../austin.ibm.com/fs/MCVS/CVS' stat for
Richard> '/.../austin.ibm.com/MCVS/CVS' stat for
Richard> '/.../MCVS/CVS' stat for '/MCVS/CVS'
Richard> 
Richard> This is for a single open.  I removed some (many)
Richard> lines.
Richard> 
Richard> The killer are these two stats:
Richard> 
Richard> stat for '/.../MCVS/CVS' stat for
Richard> '/.../{arch}/=tagging-method'
Richard> 
Richard> With IBM's dfs set up, these each hang for quite
Richard> some time (minutes).
Richard> 
Richard> I discovered vc-ignore-dir-regexp
Richard> 
Richard> I changed it to be this:
Richard> 
Richard> ## this is from my custom-set-variables in my
Richard> init.el file (vc-ignore-dir-regexp
Richard> "\\`\\([\\/][\\/]\\|/\\.\\.\\./\\|/net/\\|/afs/\\)\\'")
Richard> 
Richard> As you can see, I added /.../ into the pattern.
Richard> This seems to be exactly what I need.  Now the stats
Richard> look like this:
Richard> 
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/aix53H/{arch}/=tagging-method'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/{arch}/=tagging-method'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/{arch}/=tagging-method'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/{arch}/=tagging-method' stat
Richard> for '/.../austin.ibm.com/{arch}/=tagging-method'
Richard> ...snip...  stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/aix53H/MCVS/CVS'
Richard> stat for
Richard> '/.../austin.ibm.com/fs/projects/aix/MCVS/CVS' stat
Richard> for '/.../austin.ibm.com/fs/projects/MCVS/CVS' stat
Richard> for '/.../austin.ibm.com/fs/MCVS/CVS' stat for
Richard> '/.../austin.ibm.com/MCVS/CVS'
Richard> 
Richard> I thought this might be a good general change to add
Richard> in.  The pattern already has afs in the list.  I
Richard> believe this is the dfs equivlalent.  Although I
Richard> have not been bit by it yet, I might need to add in
Richard> /gsa/ as well for the gsa file system.  It has the
Richard> same pattern:
Richard> 
Richard> /gsa/region/path/to/file
Richard> 
Richard> So, probably stat of /gsa/CVS/ is going to hang
Richard> (although from the command line it does not seem
Richard> to).
Richard> 
Richard> I hope this makes it to the list.  I don't know if I
Richard> am signed up from this email address or not.
Richard> 
Richard> Take care, Perry Smith
Richard> 
Richard> 
Richard> _______________________________________________
Richard> emacs-pretest-bug mailing list
Richard> address@hidden
Richard> http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug
Richard> ------- End of forwarded message -------
Richard> 
Richard> 
Richard> _______________________________________________
Richard> Emacs-devel mailing list address@hidden
Richard> http://lists.gnu.org/mailman/listinfo/emacs-devel

> -- 
> Best Regards,
> --raman

      
> Email:  address@hidden
> WWW:    http://emacspeak.sf.net/raman/
> AIM:    emacspeak       GTalk: address@hidden
> PGP:    http://emacspeak.sf.net/raman/raman-almaden.asc
> Google: tv+raman 
> IRC:    irc://irc.freenode.net/#emacs


> _______________________________________________
> Emacs-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/emacs-devel