"." in SVN/entries

[Greg Stein <gs...@lyra.org>]

Dunno how it happened. I think maybe from an ra_dav commit (need to test
further), but I ended up with an entries in SVN/entries that read:

<entry
   name="."/>

Well... "svn stat" then blew up beautifully when it "recursed" into "." and
repeated the process. Over and over and over... :-)

Anyways... I suspect the ra_dav commit because I just did a commit without a
crash (found a big time pool violation). And then I went to do an svn stat
to see what the WC looked like (did ra_dav properly describe the committed
changes?). That started looping...

Not sure if the whole thing is important, but this is a heads up that we
probably aren't treating "" and "." and whatever as equivalents.

More soon...

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/

Re: "." in SVN/entries

[Ben Collins-Sussman <su...@collab.net>]

cmpilato@collab.net writes:

> I'd almost bet that there are places in the code where THIS_DIR is
> presumed to be "".  I hope I'm wrong.

You're not wrong.  I've seen them.  I've even written some of them.

(Ben begs for mercy before the firing squad...)

Re: "." in SVN/entries

[cm...@collab.net]

Greg Stein <gs...@lyra.org> writes:

> Not sure if the whole thing is important, but this is a heads up that we
> probably aren't treating "" and "." and whatever as equivalents.

Oh, dude, I think there are more issues with paths than just that.
For example, sometime in the near future I want to see what will blow
all to hell when I #define SVN_WC_ENTRY_THIS_DIR to
"IfYouSeeThisEmailCmpilatoAtCollabDotNet".  I'd almost bet that there
are places in the code where THIS_DIR is presumed to be "".  I hope
I'm wrong.

Pool Review, Path Review, String_T Review ... keep 'em coming.