| Message ID | 20221114213713.2341945-1-jonathantanmy@google.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | e62f779ae67ce3babe9e4ee934469993a5e6df49 |
| Headers | show |
| Series | [v2] Doc: document push.recurseSubmodules=only | expand |
On Mon, Nov 14, 2022 at 01:37:12PM -0800, Jonathan Tan wrote: > Documentation/config/push.txt | 14 ++-------- > Documentation/git-push.txt | 6 +++- > builtin/push.c | 12 ++++++-- > submodule.c | 6 ++++ > t/t5531-deep-submodule-push.sh | 50 ++++++++++++++++++++++++++++++++++ > 5 files changed, 73 insertions(+), 15 deletions(-) Seems reasonable to me. Glen? Thanks, Taylor
Jonathan Tan <jonathantanmy@google.com> writes: > - For me, the most important thing is to explicitly support (with > documentation) the push.recurseSubmodules=only configuration. > - Between the submodule being able to decide for itself whether > recursion continues, and the superproject deciding everything, Taylor > thinks the former is more convincing. > - The former is also the status quo. > - Glen makes a further point that either scheme (the status quo and the > behavior in v1) would not be the ideal behavior, but that a CLI > argument overrides all config, and in the absence of a CLI argument, > a submodule should be able to decide for itself whether recursion > should continue (except when there is no config present, in which case > it should defer to the superproject). Ok. I find the status quo argument convincing, so I'm okay with the approach in this patch. That said.. I find the status quo very surprising and unergonomic. As demonstrated in a test below, if a submodule has push.recurseSubmodules unset, recursion doesn't occur. At the very least, I would have expected the submodule to respect the superproject's config since that's still a statement of intent. I expect that we will have to change this behavior at some point, and hopefully my "ideal" behavior makes sense to users. > diff --git a/Documentation/config/push.txt b/Documentation/config/push.txt > index 7386fea225..43338b65e8 100644 > --- a/Documentation/config/push.txt > +++ b/Documentation/config/push.txt > @@ -110,18 +110,8 @@ This will result in only b (a and c are cleared). > ---- > > push.recurseSubmodules:: > - Make sure all submodule commits used by the revisions to be pushed > - are available on a remote-tracking branch. If the value is 'check' > - then Git will verify that all submodule commits that changed in the > - revisions to be pushed are available on at least one remote of the > - submodule. If any commits are missing, the push will be aborted and > - exit with non-zero status. If the value is 'on-demand' then all > - submodules that changed in the revisions to be pushed will be > - pushed. If on-demand was not able to push all necessary revisions > - it will also be aborted and exit with non-zero status. If the value > - is 'no' then default behavior of ignoring submodules when pushing > - is retained. You may override this configuration at time of push by > - specifying '--recurse-submodules=check|on-demand|no'. > + May be "check", "on-demand", "only", or "no", with the same behavior > + as that of "push --recurse-submodules". > If not set, 'no' is used by default, unless 'submodule.recurse' is > set (in which case a 'true' value means 'on-demand'). > > diff --git a/Documentation/git-push.txt b/Documentation/git-push.txt > index def7657ef9..5bb1d5aae2 100644 > --- a/Documentation/git-push.txt > +++ b/Documentation/git-push.txt > @@ -409,10 +409,14 @@ Specifying `--no-force-if-includes` disables this behavior. > all submodules that changed in the revisions to be pushed will be > pushed. If on-demand was not able to push all necessary revisions it will > also be aborted and exit with non-zero status. If 'only' is used all > - submodules will be recursively pushed while the superproject is left > + submodules will be pushed while the superproject is left > unpushed. A value of 'no' or using `--no-recurse-submodules` can be used > to override the push.recurseSubmodules configuration variable when no > submodule recursion is required. > ++ > +When using 'on-demand' or 'only', if a submodule has a > +"push.recurseSubmodules={on-demand,only}" or "submodule.recurse" configuration, > +further recursion will occur. In this case, "only" is treated as "on-demand". > > --[no-]verify:: > Toggle the pre-push hook (see linkgit:githooks[5]). The Everything up to here looks good. > diff --git a/builtin/push.c b/builtin/push.c > index f0329c62a2..60ac8017e5 100644 > --- a/builtin/push.c > +++ b/builtin/push.c > @@ -466,8 +466,16 @@ static int option_parse_recurse_submodules(const struct option *opt, > > if (unset) > *recurse_submodules = RECURSE_SUBMODULES_OFF; > - else > - *recurse_submodules = parse_push_recurse_submodules_arg(opt->long_name, arg); > + else { > + if (!strcmp(arg, "only-is-on-demand")) { > + if (*recurse_submodules == RECURSE_SUBMODULES_ONLY) { > + warning(_("recursing into submodule with push.recurseSubmodules=only; using on-demand instead")); > + *recurse_submodules = RECURSE_SUBMODULES_ON_DEMAND; > + } > + } else { > + *recurse_submodules = parse_push_recurse_submodules_arg(opt->long_name, arg); > + } > + } > > return 0; > } > diff --git a/submodule.c b/submodule.c > index bf7a2c7918..5cd21252d8 100644 > --- a/submodule.c > +++ b/submodule.c > @@ -1130,6 +1130,12 @@ static int push_submodule(const char *path, > if (for_each_remote_ref_submodule(path, has_remote, NULL) > 0) { > struct child_process cp = CHILD_PROCESS_INIT; > strvec_push(&cp.args, "push"); > + /* > + * When recursing into a submodule, treat any "only" configurations as "on- > + * demand", since "only" would not work (we need all submodules to be pushed > + * in order to be able to push the superproject). > + */ > + strvec_push(&cp.args, "--recurse-submodules=only-is-on-demand"); > if (dry_run) > strvec_push(&cp.args, "--dry-run"); When we pass this magic, undocumented value, "git push" will warn about about "only" and override it with "on-demand". We always pass it when we recurse into submodules, and we assume that no user will pass it, thus we get the warning iff we are recursing into submodules. In that case, it sounds like "--recurse-submodules=only-is-on-demand" is a synonym for "this is a submodule that is being recursed into". In that case, wouldn't it be more self-documenting to have a hidden CLI flag that expresses exactly that ? e.g. we could add a PARSE_OPT_HIDDEN flag called "--is-recursing" and check that boolean value. This seems clearer to me at least. > > diff --git a/t/t5531-deep-submodule-push.sh b/t/t5531-deep-submodule-push.sh > index 3f58b515ce..302e4cbdba 100755 > --- a/t/t5531-deep-submodule-push.sh > +++ b/t/t5531-deep-submodule-push.sh > @@ -512,6 +512,56 @@ test_expect_success 'push only unpushed submodules recursively' ' > test_cmp expected_pub actual_pub > ' > > +setup_subsub () { > + git init upstream && > + git init upstream/sub && > + git init upstream/sub/deepsub && > + test_commit -C upstream/sub/deepsub innermost && > + git -C upstream/sub submodule add ./deepsub deepsub && > + git -C upstream/sub commit -m middle && > + git -C upstream submodule add ./sub sub && > + git -C upstream commit -m outermost && > + > + git -c protocol.file.allow=always clone --recurse-submodules upstream downstream && > + git -C downstream/sub/deepsub checkout -b downstream-branch && > + git -C downstream/sub checkout -b downstream-branch && > + git -C downstream checkout -b downstream-branch > +} > + > +new_downstream_commits () { > + test_commit -C downstream/sub/deepsub new-innermost && > + git -C downstream/sub add deepsub && > + git -C downstream/sub commit -m new-middle && > + git -C downstream add sub && > + git -C downstream commit -m new-outermost > +} > + > +test_expect_success 'push with push.recurseSubmodules=only on superproject' ' > + test_when_finished rm -rf upstream downstream && > + setup_subsub && > + new_downstream_commits && > + git -C downstream config push.recurseSubmodules only && > + git -C downstream push origin downstream-branch && > + > + test_must_fail git -C upstream rev-parse refs/heads/downstream-branch && > + git -C upstream/sub rev-parse refs/heads/downstream-branch && > + test_must_fail git -C upstream/sub/deepsub rev-parse refs/heads/downstream-branch > +' This test demonstrates the behavior I find surprising. > + > +test_expect_success 'push with push.recurseSubmodules=only on superproject and top-level submodule' ' > + test_when_finished rm -rf upstream downstream && > + setup_subsub && > + new_downstream_commits && > + git -C downstream config push.recurseSubmodules only && > + git -C downstream/sub config push.recurseSubmodules only && > + git -C downstream push origin downstream-branch 2> err && > + > + test_must_fail git -C upstream rev-parse refs/heads/downstream-branch && > + git -C upstream/sub rev-parse refs/heads/downstream-branch && > + git -C upstream/sub/deepsub rev-parse refs/heads/downstream-branch && > + grep "recursing into submodule with push.recurseSubmodules=only; using on-demand instead" err > +' > + > test_expect_success 'push propagating the remotes name to a submodule' ' > git -C work remote add origin ../pub.git && > git -C work remote add pub ../pub.git && The tests look good, and are quite readable. > -- > 2.38.1.493.g58b659f92b-goog
First of all, thanks for taking a look at this. Glen Choo <chooglen@google.com> writes: > That said.. I find the status quo very surprising and unergonomic. As > demonstrated in a test below, if a submodule has push.recurseSubmodules > unset, recursion doesn't occur. At the very least, I would have expected > the submodule to respect the superproject's config since that's still a > statement of intent. > > I expect that we will have to change this behavior at some point, and > hopefully my "ideal" behavior makes sense to users. Well, your "ideal" behavior makes sense to me :-) > When we pass this magic, undocumented value, "git push" will warn about > about "only" and override it with "on-demand". We always pass it when we > recurse into submodules, and we assume that no user will pass it, thus > we get the warning iff we are recursing into submodules. > > In that case, it sounds like "--recurse-submodules=only-is-on-demand" is > a synonym for "this is a submodule that is being recursed into". In that > case, wouldn't it be more self-documenting to have a hidden CLI flag > that expresses exactly that ? e.g. we could add a PARSE_OPT_HIDDEN flag > called "--is-recursing" and check that boolean value. This seems clearer > to me at least. Hmm...--recurse-submodules=only-is-on-demand is hidden too, right? One advantage of doing this instead of a separate arg is that neither the caller nor "git push" needs to think about what happens if both --recurse- submodules=something and --is-recursing are both specified.
Jonathan Tan <jonathantanmy@google.com> writes: >> When we pass this magic, undocumented value, "git push" will warn about >> about "only" and override it with "on-demand". We always pass it when we >> recurse into submodules, and we assume that no user will pass it, thus >> we get the warning iff we are recursing into submodules. >> >> In that case, it sounds like "--recurse-submodules=only-is-on-demand" is >> a synonym for "this is a submodule that is being recursed into". In that >> case, wouldn't it be more self-documenting to have a hidden CLI flag >> that expresses exactly that ? e.g. we could add a PARSE_OPT_HIDDEN flag >> called "--is-recursing" and check that boolean value. This seems clearer >> to me at least. > > Hmm...--recurse-submodules=only-is-on-demand is hidden too, right? I suppose so, but as a matter of personal taste, when encountering a hidden CLI option, I'd prefer to see the option's documentation when I grep: OPT_HIDDEN_BOOL(0, "is-recursing", &is_recursing, "internal only, override recurseSubmodules = only") than: if (!strcmp(arg, "only-is-on-demand")) { /* etc */ } > One > advantage of doing this instead of a separate arg is that neither the > caller nor "git push" needs to think about what happens if both --recurse- > submodules=something and --is-recursing are both specified. That's true, and I suppose it prevents the temptation to reuse --is-recursing for things it wasn't meant for. But in that case, couldn't we say the same thing if we renamed "--is-recursing" to "--recurse-only-is-on-demand"? At any rate, do treat this as just a suggestion. I won't block this if both you and Taylor find this clear enough :)
On Wed, Nov 16, 2022 at 01:21:03PM -0800, Glen Choo wrote: > At any rate, do treat this as just a suggestion. I won't block this if > both you and Taylor find this clear enough :) I admit that I do not care too much either way ;-). Thanks, Taylor
diff --git a/Documentation/config/push.txt b/Documentation/config/push.txt index 7386fea225..43338b65e8 100644 --- a/Documentation/config/push.txt +++ b/Documentation/config/push.txt @@ -110,18 +110,8 @@ This will result in only b (a and c are cleared). ---- push.recurseSubmodules:: - Make sure all submodule commits used by the revisions to be pushed - are available on a remote-tracking branch. If the value is 'check' - then Git will verify that all submodule commits that changed in the - revisions to be pushed are available on at least one remote of the - submodule. If any commits are missing, the push will be aborted and - exit with non-zero status. If the value is 'on-demand' then all - submodules that changed in the revisions to be pushed will be - pushed. If on-demand was not able to push all necessary revisions - it will also be aborted and exit with non-zero status. If the value - is 'no' then default behavior of ignoring submodules when pushing - is retained. You may override this configuration at time of push by - specifying '--recurse-submodules=check|on-demand|no'. + May be "check", "on-demand", "only", or "no", with the same behavior + as that of "push --recurse-submodules". If not set, 'no' is used by default, unless 'submodule.recurse' is set (in which case a 'true' value means 'on-demand'). diff --git a/Documentation/git-push.txt b/Documentation/git-push.txt index def7657ef9..5bb1d5aae2 100644 --- a/Documentation/git-push.txt +++ b/Documentation/git-push.txt @@ -409,10 +409,14 @@ Specifying `--no-force-if-includes` disables this behavior. all submodules that changed in the revisions to be pushed will be pushed. If on-demand was not able to push all necessary revisions it will also be aborted and exit with non-zero status. If 'only' is used all - submodules will be recursively pushed while the superproject is left + submodules will be pushed while the superproject is left unpushed. A value of 'no' or using `--no-recurse-submodules` can be used to override the push.recurseSubmodules configuration variable when no submodule recursion is required. ++ +When using 'on-demand' or 'only', if a submodule has a +"push.recurseSubmodules={on-demand,only}" or "submodule.recurse" configuration, +further recursion will occur. In this case, "only" is treated as "on-demand". --[no-]verify:: Toggle the pre-push hook (see linkgit:githooks[5]). The diff --git a/builtin/push.c b/builtin/push.c index f0329c62a2..60ac8017e5 100644 --- a/builtin/push.c +++ b/builtin/push.c @@ -466,8 +466,16 @@ static int option_parse_recurse_submodules(const struct option *opt, if (unset) *recurse_submodules = RECURSE_SUBMODULES_OFF; - else - *recurse_submodules = parse_push_recurse_submodules_arg(opt->long_name, arg); + else { + if (!strcmp(arg, "only-is-on-demand")) { + if (*recurse_submodules == RECURSE_SUBMODULES_ONLY) { + warning(_("recursing into submodule with push.recurseSubmodules=only; using on-demand instead")); + *recurse_submodules = RECURSE_SUBMODULES_ON_DEMAND; + } + } else { + *recurse_submodules = parse_push_recurse_submodules_arg(opt->long_name, arg); + } + } return 0; } diff --git a/submodule.c b/submodule.c index bf7a2c7918..5cd21252d8 100644 --- a/submodule.c +++ b/submodule.c @@ -1130,6 +1130,12 @@ static int push_submodule(const char *path, if (for_each_remote_ref_submodule(path, has_remote, NULL) > 0) { struct child_process cp = CHILD_PROCESS_INIT; strvec_push(&cp.args, "push"); + /* + * When recursing into a submodule, treat any "only" configurations as "on- + * demand", since "only" would not work (we need all submodules to be pushed + * in order to be able to push the superproject). + */ + strvec_push(&cp.args, "--recurse-submodules=only-is-on-demand"); if (dry_run) strvec_push(&cp.args, "--dry-run"); diff --git a/t/t5531-deep-submodule-push.sh b/t/t5531-deep-submodule-push.sh index 3f58b515ce..302e4cbdba 100755 --- a/t/t5531-deep-submodule-push.sh +++ b/t/t5531-deep-submodule-push.sh @@ -512,6 +512,56 @@ test_expect_success 'push only unpushed submodules recursively' ' test_cmp expected_pub actual_pub ' +setup_subsub () { + git init upstream && + git init upstream/sub && + git init upstream/sub/deepsub && + test_commit -C upstream/sub/deepsub innermost && + git -C upstream/sub submodule add ./deepsub deepsub && + git -C upstream/sub commit -m middle && + git -C upstream submodule add ./sub sub && + git -C upstream commit -m outermost && + + git -c protocol.file.allow=always clone --recurse-submodules upstream downstream && + git -C downstream/sub/deepsub checkout -b downstream-branch && + git -C downstream/sub checkout -b downstream-branch && + git -C downstream checkout -b downstream-branch +} + +new_downstream_commits () { + test_commit -C downstream/sub/deepsub new-innermost && + git -C downstream/sub add deepsub && + git -C downstream/sub commit -m new-middle && + git -C downstream add sub && + git -C downstream commit -m new-outermost +} + +test_expect_success 'push with push.recurseSubmodules=only on superproject' ' + test_when_finished rm -rf upstream downstream && + setup_subsub && + new_downstream_commits && + git -C downstream config push.recurseSubmodules only && + git -C downstream push origin downstream-branch && + + test_must_fail git -C upstream rev-parse refs/heads/downstream-branch && + git -C upstream/sub rev-parse refs/heads/downstream-branch && + test_must_fail git -C upstream/sub/deepsub rev-parse refs/heads/downstream-branch +' + +test_expect_success 'push with push.recurseSubmodules=only on superproject and top-level submodule' ' + test_when_finished rm -rf upstream downstream && + setup_subsub && + new_downstream_commits && + git -C downstream config push.recurseSubmodules only && + git -C downstream/sub config push.recurseSubmodules only && + git -C downstream push origin downstream-branch 2> err && + + test_must_fail git -C upstream rev-parse refs/heads/downstream-branch && + git -C upstream/sub rev-parse refs/heads/downstream-branch && + git -C upstream/sub/deepsub rev-parse refs/heads/downstream-branch && + grep "recursing into submodule with push.recurseSubmodules=only; using on-demand instead" err +' + test_expect_success 'push propagating the remotes name to a submodule' ' git -C work remote add origin ../pub.git && git -C work remote add pub ../pub.git &&
Git learned pushing submodules without pushing the superproject by the user specifying --recurse-submodules=only through 6c656c3fe4 ("submodules: add RECURSE_SUBMODULES_ONLY value", 2016-12-20) and 225e8bf778 ("push: add option to push only submodules", 2016-12-20). For users who use this feature regularly, it is desirable to have an equivalent configuration. It turns out that such a configuration (push.recurseSubmodules=only) is already supported, even though it is neither documented nor mentioned in the commit messages, due to the way the --recurse-submodules=only feature was implemented (a function used to parse --recurse-submodules was updated to support "only", but that same function is used to parse push.recurseSubmodules too). What is left is to document it and test it, which is what this commit does. There is a possible point of confusion when recursing into a submodule that itself has the push.recurseSubmodules=only configuration, because if a repository has only its submodules pushed and not itself, its superproject can never be pushed. Therefore, treat such configurations as being "on-demand", and print a warning message. Signed-off-by: Jonathan Tan <jonathantanmy@google.com> --- Thanks once again, Taylor and Glen for your comments. To try to summarize: - For me, the most important thing is to explicitly support (with documentation) the push.recurseSubmodules=only configuration. - Between the submodule being able to decide for itself whether recursion continues, and the superproject deciding everything, Taylor thinks the former is more convincing. - The former is also the status quo. - Glen makes a further point that either scheme (the status quo and the behavior in v1) would not be the ideal behavior, but that a CLI argument overrides all config, and in the absence of a CLI argument, a submodule should be able to decide for itself whether recursion should continue (except when there is no config present, in which case it should defer to the superproject). If we were to implement v1 and then the ideal behavior later, we would need to revert some of the v1-introduced behavior: if there is configuration in both the superproject and a submodule, the status quo is that the submodule's configuration takes precedence, v1 has the superproject's configuration taking precedence, and with ideal behavior, the submodule's configuration takes precedence. So it makes sense to stick to the status quo for this patch. It is possible to support push.recurseSubmodules=only with all existing behavior preserved (in fact, push.recurseSubmodules=only is already supported, just undocumented) so I have decided to go with that approach. --- Documentation/config/push.txt | 14 ++-------- Documentation/git-push.txt | 6 +++- builtin/push.c | 12 ++++++-- submodule.c | 6 ++++ t/t5531-deep-submodule-push.sh | 50 ++++++++++++++++++++++++++++++++++ 5 files changed, 73 insertions(+), 15 deletions(-)