On Sat, Mar 16, 2013 at 11:40 PM, Sami Kerola <kerolasa@xxxxxx> wrote:
> Here are the results of the whereis(1) work which I mentioned earlier the
> week in irc. All of the patches are related to that command, and as you
> can see from the diff stats a lot has changed.
>
> The first patch is quite big. Unfortunately I ran out of imagination how
> to make it smaller, as the patch is the core change enabling rest of the
> work. Most what the 0001 has got going is with search directory list
> building using new data structure.
>
> In between patches are quite straight forward till second last which will
> canonicalize output file names. I think that is what users are
> interested, but I a may be wrong as well. Comments, anyone?
>
> The last patch is a bug fix to RedHat item. See commit message for
> further details.
>
> One funky side note. After the changes the whereis seems to run about
> 20% quicker. That was unintentional enhancement.
Hi,
Is this regression, or a command behavior change resulting to correct
output when not patched? I think earlier whereis misbehaved. Reason
being is that shell 'which' or path completion does not recognize
broken symlink.
$ cd ~/bin
$ ln -s bar foo
$ whereis foo
foo: /home/kerolasa/bin/foo
$ cd ~/src/util-linux
$ ./whereis foo
# no results
$ git diff
diff --git a/misc-utils/whereis.c b/misc-utils/whereis.c
index 0e333e1..30a6819 100644
--- a/misc-utils/whereis.c
+++ b/misc-utils/whereis.c
@@ -179,7 +179,8 @@ static int is_same_file(char *file, struct is_here
*filelist)
struct is_here *prev = NULL;
if (stat(file, &statbuf))
- return 1;
+ if (lstat(file, &statbuf))
+ return 1;
while (filelist) {
if (filelist->st_ino == statbuf.st_ino &&
filelist->st_dev == statbuf.st_dev)
$ make
# <...>
./whereis foo
foo: /home/kerolasa/bin/foo
--
Sami Kerola
http://www.iki.fi/kerolasa/
--
To unsubscribe from this list: send the line "unsubscribe util-linux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
