Re: maint.mk: public-submodule-commit rule is broken
On Mon, Sep 12, 2022 at 3:32 AM Simon Josefsson wrote: > Btw, see my concerns with this code earlier here: > > https://lists.gnu.org/archive/html/bug-gnulib/2022-08/msg00040.html > https://lists.gnu.org/archive/html/bug-gnulib/2022-08/msg00044.html > > Instead of the patch, I have merely disabled it when I ran into issues, > since it doesn't add value for me and causes problems. > > Couldn't this be converted into a syntax-check rule instead? Or just > removed. Hi Bruno and Simon, Sorry it's causing trouble. You're welcome to improve, delete or move it to a syntax-check rule. I prefer the latter, because there it may still help someone, and it's easy to disable.
Re: maint.mk: public-submodule-commit rule is broken
Btw, see my concerns with this code earlier here: https://lists.gnu.org/archive/html/bug-gnulib/2022-08/msg00040.html https://lists.gnu.org/archive/html/bug-gnulib/2022-08/msg00044.html Instead of the patch, I have merely disabled it when I ran into issues, since it doesn't add value for me and causes problems. Couldn't this be converted into a syntax-check rule instead? Or just removed. /Simon signature.asc Description: PGP signature
maint.mk: public-submodule-commit rule is broken
Hi Jim,
In GNU gettext, I'm now using one of the 'stable-*' branches of gnulib.
And maint.mk gives me an error:
$ make check
if test -d ./.git \
&& git --version >/dev/null 2>&1; then \
cd . && \
git submodule --quiet foreach \
'test "$(git rev-parse "$sha1")" \
= "$(git merge-base origin "$sha1")"' \
|| { echo 'maint.mk: found non-public submodule commit' >&2;\
exit 1; }; \
else\
: ; \
fi
fatal: run_command returned non-zero status for gnulib
.
maint.mk: found non-public submodule commit
make: *** [maint.mk:1489: public-submodule-commit] Error 1
Two things are wrong here:
1) I should not be getting a "found non-public submodule commit" error,
since the gnulib commit that I am using *is* public. See:
$ git submodule foreach 'echo $sha1; test "$(git rev-parse "$sha1")" = "$(git
merge-base origin "$sha1")"'
Entering 'gnulib'
dc631883e47212d5cd311575b491cb5b093b34d9
fatal: run_command returned non-zero status for gnulib
.
$ git rev-parse dc631883e47212d5cd311575b491cb5b093b34d9
dc631883e47212d5cd311575b491cb5b093b34d9
$ git merge-base origin dc631883e47212d5cd311575b491cb5b093b34d9
fatal: Not a valid commit name dc631883e47212d5cd311575b491cb5b093b34d9
So,
git merge-base origin "$sha1"
fails. But that does not mean that the commit is non-public. You can see
the commit at
https://git.savannah.gnu.org/gitweb/?p=gnulib.git;a=commit;h=dc631883e47212d5cd311575b491cb5b093b34d9
2) The 'public-submodule-commit' is not suitable for hooking into
"make check". It would be better hooked into "make distcheck".
Why? Because as a developer who has write access to a submodule the normal
workflow is like this:
1. Check out a public master or branch of the submodule.
2. Prepare a change to the submodule. Commit it but don't push it.
3. In the parent package, do "make" and "make check", to verify that it
doesn't cause a regression.
4. Push the change in the submodule.
If you hook 'public-submodule-commit' into "make check", you are forcing
the developer to swap steps 3 and 4. That is, to push a change to the public,
before having verified that it does not cause regressions!
Bruno
