> From: Junio C Hamano [mailto:[email protected]]
> Sent: Sunday, August 19, 2012 7:23 PM
> To: Joachim Schmitz
> Cc: 'Shawn Pearce'; [email protected]; [email protected]
> Subject: Re: Porting git to HP NonStop
>
> "Joachim Schmitz" <[email protected]> writes:
>
> > Found the problem: our mkdir(dir,flags) fails with ENOENT when dir
> > ends with a '/'.
> > Not sure whether this us a bug on out platform or just allowed by
> > POSIX and as such a wrong assumption in git though?
> >
> > [shortly after]
> > A bit of googleing revealed that there is a GNUlib solution for this,
> > which claims that at least NetBSD 1.5.2 has the same problem.
> > (http://www.opensource.apple.com/source/gpatch/gpatch-2/patch/mkdir.c)
> >
> > And apparently this has been discussed on the git mailing list too, 2
> > years
> > ago:
> > http://lists-archives.com/git/728359-git-s-use-of-mkdir-2.html,
> > there's a patch too.
>
> Given that newer BSDs have fixed libc to accept directory name with a
trailing
> slash, and that we use mkdir(2) in many places, I think the right way to
do so is
> still what I suggested in that old thread in the last paragraph of my
message
>
> http://thread.gmane.org/gmane.comp.version-
> control.git/155812/focus=155876
>
> That is, have compat/tandem.c and define a replacement mkdir(2) in a way
> similar to how MinGW does so.
OK, I'll go for a compat/mkdir.c though.
We shouldn't call it tandem.c as Tandem, the Company, doesn't exist anymore
and since more than a decade (bough by Compaq, then HP), only the __TANDEM
survived in our compiler and headers/libraries. Could call it NonStop.c, but
I don't really like that idea either, I'd rather keep it more generic, just
in case someone else might need it too, or that issue someday gets fixed for
NonStop.
> > For now I've fixed it like this:
> > /usr/local/bin/diff -EBbu ./builtin/init-db.c.orig ./builtin/init-db.c
> > --- ./builtin/init-db.c.orig 2012-08-19 03:55:50 -0500
> > +++ ./builtin/init-db.c 2012-08-19 03:39:57 -0500
> > @@ -25,7 +25,16 @@
> >
> > static void safe_create_dir(const char *dir, int share) {
> > +#ifdef __TANDEM /* our mkdir() can't cope with a trailing '/' */
> > + char mydir[PATH_MAX];
> > +
> > + strcpy(mydir,dir);
> > + if (dir[strlen(dir)-1] == '/')
> > + mydir[strlen(dir)-1] = '\0';
> > + if (mkdir(mydir, 0777) < 0) {
> > +#else
> > if (mkdir(dir, 0777) < 0) {
> > +#endif
>
> Move that part inside #ifdef __TANDEM to define
>
> int tandem_mkdir(const char *dir, mode_t mode)
> {
> ...
> }
I'll go for git_mkdir(), similar to other git wrappers, (like for mmap,
pread, fopen, snprintf, vsnprintf, qsort). Could call it gitmkdir() too
(like for basename, setenv, mkdtemp, mkstemps, unsetenv, strcasestr,
strlcpy, strtoumax, strtoimax, strtok_r, hstrerror, memmem, strchrnul,
memcpy), Opinions?
It seems the ones without the "_" are for missing APIs and the ones with "_"
to wrap existing APIs (not sure about mmap and pread)?
Here it's current state:
$ cat compat/mkdir.c
#include "../git-compat-util.h"
#undef mkdir
/* for platforms that can't deal with a trailing '/' */
int git_mkdir(const char *dir, mode_t mode)
{
int retval;
char *tmp_dir = NULL;
size_t len = strlen(dir);
if (len && dir[len-1] == '/') {
if ((tmp_dir = strdup(dir)) == NULL)
return -1;
tmp_dir[len-1] = '\0';
}
else
tmp_dir = (char *)dir;
retval = mkdir(tmp_dir, mode);
if (tmp_dir != dir)
free(tmp_dir);
return retval;
}
$
There is room for improvement though: it only removes one trailing slash. By
far not as advanced and generic as GNUlib's mkdir wrapper, but should be
good enough for git's usage.
> in your new file compat/tandem.c, add
>
> #ifdef __TANDEM
> #define mkdir(a,b) tandem_mkdir((a), (b))
> #endif
>
> to git-compat-util.h
Again, git_mkdir, see above
> and then add compat/tandem.o to COMPAT_OBJS in the
> top-level Makefile.
For now I've added it to the (new) NOSTOP_KERNEL section.
We may want it to go along with some
MKDIR_DISLIKES_TRAILING_SLASH or MKDIR_BOGUS_TRAILING_SLASH some such.
Opinions, Ideas?
> That way we do not have to keep an ugly platform specific ifdef in the
very
> generic codepath.
Agreed, it was my quick and dirty fix for it.
Bye, Jojo
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
