[PATCH] Fix util._statfiles_clustered() failing at root of a windows drive

[Petr Kodl]

how about this - sorry for late entry

diff -r c4461ea8b4c8 mercurial/util.py
--- a/mercurial/util.py	Sun Oct 26 17:26:28 2008 +0100
+++ b/mercurial/util.py	Sat Nov 01 12:01:55 2008 -0400
@@ -856,7 +856,7 @@
          if pos == -1:
              dir, base = '.', nf
          else:
-            dir, base = nf[:pos], nf[pos+1:]
+            dir, base = nf[:pos+1], nf[pos+1:]
          cache = dircache.get(dir, None)
          if cache is None:
              try:
@@ -874,7 +874,7 @@
              cache = dircache.setdefault(dir, dmap)
          yield cache.get(base, None)

This leaves the trailing \ in dir so m:\ stays m:\ instead of becoming m:

osutil.c can handle the trailing slash just fine - it turns it into \*.* 
and this behavior is consistent in native Python os.listdir

I tested both root and non-root repos and it works fine

pk



Patrick Mézard wrote:
> Patrick Mézard a écrit :
>> Any objection about the following instead ? Why aren't we keeping the trailing os.sep in dir, I don't know.
> 
> Keeping the trailing os.sep does not seem problematic, what about this one instead:
> 
> diff --git a/mercurial/util.py b/mercurial/util.py
> --- a/mercurial/util.py
> +++ b/mercurial/util.py
> @@ -849,6 +849,7 @@
>      lstat = os.lstat
>      ncase = os.path.normcase
>      sep   = os.sep
> +    rootpath = None
>      dircache = {} # dirname -> filename -> status | None if file does not exist
>      for nf in files:
>          nf  = ncase(nf)
> @@ -856,7 +857,11 @@
>          if pos == -1:
>              dir, base = '.', nf
>          else:
> -            dir, base = nf[:pos], nf[pos+1:]
> +            dir, base = nf[:pos+1], nf[pos+1:]
> +            if dir == sep:
> +                if rootpath is None:
> +                    rootpath = os.path.abspath(sep)
> +                dir = rootpath
>          cache = dircache.get(dir, None)
>          if cache is None:
>              try:
> 
> _______________________________________________
> Mercurial-devel mailing list
> Mercurial-devel at selenic.com
> http://selenic.com/mailman/listinfo/mercurial-devel
>