Re: Teach git-update-index about gitlinks

From: Alex Riesen <raa.lkml@gmail.com>
Date: Apr 13, 2007, 07:42 UTC
Message-ID: <81b0412b0704130042k1f5be3bay30cbbb6b70259f6@mail.gmail.com>
In-Reply-To: <Pine.LNX.4.64.0704121218130.4061@woody.linux-foundation.org>

On 4/12/07, Linus Torvalds <torvalds@linux-foundation.org> wrote:

Show 5 quoted lines

> +static int process_lstat_error(const char *path, int err)
> +{
> +       if (err == ENOENT || err == ENOTDIR)
> +               return remove_one_path(path);
> +       return error("lstat(\"%s\"): %s", path, strerror(errno));

You forgot to change errno to err in strerror.

> +static int add_one_path(struct cache_entry *old, const char *path, int len, struct stat *st)

...

Show 6 quoted lines

>         option = allow_add ? ADD_CACHE_OK_TO_ADD : 0;
>         option |= allow_replace ? ADD_CACHE_OK_TO_REPLACE : 0;
>         if (add_cache_entry(ce, option))
> -               return error("%s: cannot add to the index - missing --add option?",
> -                            path);
> +               return error("%s: cannot add to the index - missing --add option?", path);

return error("%s: cannot add to the index%s", path, "\0 - missing --add option?" + !allow_add); is be nicer. We really do know whether "--add" was passed or not.

> + *    (NOTE! This is old and arguably fairly strange behaviour.
> + *    We might want to make this an error unconditionally, and
> + *    use "--force-remove" if you actually want to force removal).

...

> +               /* Should this be an unconditional error? */
> +               return remove_one_path(path);

Makes me uncomfortable too. Is it be possible anyone is depending on it? Maybe it still can be changed?