Re: [PATCH] Add git-submodule command

On 5/25/07, Johannes Schindelin [off-list ref] wrote:

Hi,

On Fri, 25 May 2007, Lars Hjemli wrote:

quoted

There is currently no way to override the mappings in the .gitmodules
file, except by manually creating the subproject repository.

I think that is okay. We can add that easily at a later stage, and the
script is much easier without that logic.

Yes. And it is sort of a feature: if you've cloned the submodule using
a different (aka local) url, 'git submodule --init' will leave it
alone.

quoted

diff --git a/Documentation/git-submodule.txt b/Documentation/git-submodule.txt

Looks good here, I checked with asciidoc.

Thanks.

quoted

diff --git a/git-submodule.sh b/git-submodule.sh
new file mode 100755
index 0000000..c4a1cc3
--- /dev/null
+++ b/git-submodule.sh

@@ -0,0 +1,163 @@
+#!/bin/sh
+#
+# git-submodule.sh: init, update or list git submodules
+#
+# Copyright (c) 2007 Lars Hjemli
+
+USAGE='[-i | --init | -u | --update] [-q | --quiet] [--cached] <path>...'
+. git-sh-setup
+require_work_tree
+
+init=
+update=
+quiet=
+cached=
+
+#
+# print stuff on stdout unless -q was specified
+#
+say()
+{
+     if test -z "$quiet"
+     then
+             echo -e "$@"
+     fi
+}
+
+#
+# Find all (requested) submodules, run clone + checkout on missing paths
+#
+# $@ = requested paths (default to all)
+#
+modules_init()
+{
+     git ls-files --stage -- $@ | grep -e '^160000 ' |

Any reason you read in the stage? It does not seem that you use it.

Are there any other way to get the mode info?

quoted

+     while read mode sha1 stage path
+     do
+             test -d "$path/.git" && continue
+
+             if test -d "$path"
+             then
+                     rmdir "$path" 2>/dev/null ||
+                     die "Directory '$path' exist, but not as a submodule"
+             fi
+
+             test -e "$path" &&
+             die "A file already exist at path '$path'"
+
+             url=$(GIT_CONFIG=.gitmodules git-config module."$path".url)

I like that command ;-)

quoted

+             test -z "$url" &&
+             die "No url found for submodule '$path' in .gitmodules"
+
+             git-clone "$url" "$path" ||
+             die "Clone of submodule '$path' failed"
+
+             $(cd "$path" && git-checkout -q "$sha1") ||
+             die "Checkout of submodule '$path' failed"
+
+             say "Submodule '$path' initialized"
+     done
+}
+
+#
+# Checkout correct revision of each initialized submodule
+#
+# $@ = requested paths (default to all)
+#
+modules_update()
+{
+     git ls-files --stage -- $@ | grep -e '^160000 ' |

Same here.

quoted

+     while read mode sha1 stage path
+     do
+             if ! test -d "$path/.git"
+             then
+                     say "Submodule '$path' not initialized"
+                     continue;
+             fi
+             subsha1=$(cd "$path" && git-rev-parse --verify HEAD) ||

Maybe it would be a better idea to use "git --git-dir="$path" rev-parse
..."? Just in case somebody calls this with GIT_DIR overridden...

Or, unset GIT_DIR explicitely.

Hmm, that's annoying (overridden GIT_DIR). I guess 'git --git-dir
$path/.git' would be the easiest solution.

quoted

+             die "Unable to find current revision of submodule '$path'"
+
+             if test "$subsha1" != "$sha1"
+             then
+                     $(cd "$path" && git-fetch && git-checkout -q "$sha1") ||

This will make a detached HEAD, right? Do you want that? (I am not really
interested in submodules myself, so I haven't thought about it, and I
haven't followed that monster discussion.)

Well, we might want to be smarter about this, but on the other hand:
if the user cares, he can always do 'cd $path && git checkout
$branch', since 'git submodule -u' will skip submodules with the
correct commit checked out.

quoted

+                     die "Unable to checkout revision $sha1 of submodule '$path'"
+
+                     say "Submodule '$path' reset to revision $sha1"

I'd rather not say "reset", since this has a different meaning in Git, but
rather "set to revision $sha1".

Ok.

quoted

+             fi
+     done
+}
+
+#
+# List all registered submodules, prefixed with:
+#  - submodule not initialized
+#  + different version checked out
+#
+# If --cached was specified the revision in the index will be printed
+# instead of the currently checked out revision.
+#
+# $@ = requested paths (default to all)
+#
+modules_list()
+{
+     git ls-files --stage -- $@ | grep -e '^160000 ' |
+     while read mode sha1 stage path
+     do
+             if ! test -d "$path/.git"
+             then
+                     say "-$sha1 $path"
+                     continue;
+             fi
+             revname=$(cd "$path" && git-describe $sha1)
+             if git diff-files --quiet -- "$path"
+             then
+                     say " $sha1 $path\t($revname)"
+             else
+                     if test -z "$cached"
+                     then
+                             sha1=$(cd "$path" && git-rev-parse --verify HEAD)
+                             revname=$(cd "$path" && git-describe $sha1)
+                     fi
+                     say "+$sha1 $path\t($revname)"
+             fi
+     done
+}
+
+
+while case "$#" in 0) break ;; esac
+do
+     case "$1" in
+     -i|--init)
+             init=1
+             ;;
+     -u|--update)
+             update=1
+             ;;
+     -q|--quiet)
+             quiet=1
+             ;;
+     --cached)
+             cached=1
+             ;;
+     --)
+             break
+             ;;
+     -*)
+             usage
+             ;;
+     *)
+             break
+             ;;
+     esac
+     shift
+done
+
+
+if test "$init" = "1"
+then
+     modules_init $@
+elif test "$update" = "1"
+then
+     modules_update $@
+else
+     modules_list $@
+fi

I'll let Junio comment on that command line parsing...

Heh, I'm a shell illiterate...

All in all, I like it: it is short, to the point, and it should do the job
(maybe with a few enhancements like "--update" without arguments means
_all_ submodules).

Well, it does (or should) update all initialized submodules, but maybe
that's not what you meant?

Thanks for the review!

--
larsh