Re: [PATCH] Remove perl dependency from git-submodule.sh
- From
Fredrik Gustafsson <iveqy@iveqy.com>
- Date
- May 31, 2012, 10:40 UTC
- Message-ID
- <20120531104036.GB30500@paksenarrion.iveqy.com>
- In-Reply-To
- <4FC73788.6070805@viscovery.net>
On Thu, May 31, 2012 at 11:19:04AM +0200, Johannes Sixt wrote:
Show 35 quoted lines
> Am 5/31/2012 10:48, schrieb Fredrik Gustafsson:
> > Rewrote a perl section in sh.
>
> > The code may be a bit slower (doing grep on strings instead of using
> > perl-lists).
>
> "A lot" would be more correct on Windows :-) But it can be avoided, I think.
>
> > module_list()
> > {
> > + unmerged=
> > + null_sha1=0000000000000000000000000000000000000000
> > git ls-files --error-unmatch --stage -- "$@" |
> > - perl -e '
> > - my %unmerged = ();
> > - my ($null_sha1) = ("0" x 40);
> > - while (<STDIN>) {
> > - chomp;
> > - my ($mode, $sha1, $stage, $path) =
> > - /^([0-7]+) ([0-9a-f]{40}) ([0-3])\t(.*)$/;
> > - next unless $mode eq "160000";
> > - if ($stage ne "0") {
> > - if (!$unmerged{$path}++) {
> > - print "$mode $null_sha1 U\t$path\n";
> > - }
> > - next;
> > - }
> > - print "$_\n";
> > - }
> > - '
> > + while read mode sha1 stage path
>
> Be prepared for backslashes in the path name:
>
> while read -r mode sha1 stage pathWe are not using -r on any place in git-submodule.sh. Maybe we should? I can provide a patch if needed.
Show 5 quoted lines
> > > + do > > + if test $mode -eq 160000 > > $mode is not a number, but a string: test "$mode" = 160000
okay, fixed in next iteration.
Show 5 quoted lines
> > > + then > > + if test $stage -ne 0 > > That $stage looks like a number is of no importance, either.
Actually I don't know what stage does and if it's important here. This part is just to mimic the perl code. Should it be removed?
Show 12 quoted lines
> > > + then > > + if test -z "$(echo $unmerged | grep "|$path|")" > > + then > > + echo "$mode $null_sha1 U\t$path" > > + fi > > + unmerged="$unmerged|$path|" > > IIUC, the purpose of $unmerged and this check is to avoid that an unmerged > path is dumped for each stage that is listed by ls-files. Therefore it > should be sufficient to just check that the current path is different from > the last path.
That requires that submodules always is in the same order, right? That would not work for: sub1 sub2 sub1
case. But is that order really a possibility or is it always sub1 sub1 sub2
?
-- Med vänliga hälsningar Fredrik Gustafsson tel: 0733-608274 e-post: iveqy@iveqy.com