Sorry about the confusion. Just so I'm 100% clear: the `submodule.<name>.update` referencing a gitconfig and not gitmodules? On Thu, Nov 3, 2022 at 8:17 AM Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> wrote: > > > On Thu, Nov 03 2022, James via GitGitGadget wrote: > > > From: James <jdm12989@xxxxxxxxx> > > > > Removes reference to custom code execution which is called out by > > gitmodules as not allowed. > > > > Signed-off-by: James <jdm12989@xxxxxxxxx> > > --- > > Update git-submodule.txt > > > > Remove info about custom commands being allowed in the submodule update > > command to align with the gitmodules docs which states "For security > > reasons, the !command form is not accepted here." > > > > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1401%2FEnochPrime%2Fsubmodule-docs-v1 > > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1401/EnochPrime/submodule-docs-v1 > > Pull-Request: https://github.com/gitgitgadget/git/pull/1401 > > > > Documentation/git-submodule.txt | 6 ------ > > 1 file changed, 6 deletions(-) > > > > diff --git a/Documentation/git-submodule.txt b/Documentation/git-submodule.txt > > index 4d3ab6b9f92..b40ac72f753 100644 > > --- a/Documentation/git-submodule.txt > > +++ b/Documentation/git-submodule.txt > > @@ -163,12 +163,6 @@ checked out in the submodule. > > The following 'update' procedures are only available via the > > `submodule.<name>.update` configuration variable: > > > > - custom command;; arbitrary shell command that takes a single > > - argument (the sha1 of the commit recorded in the > > - superproject) is executed. When `submodule.<name>.update` > > - is set to '!command', the remainder after the exclamation mark > > - is the custom command. > > - > > none;; the submodule is not updated. > > > > If the submodule is not yet initialized, and you just want to use the > > Hrm? > > diff --git a/builtin/submodule--helper.c b/builtin/submodule--helper.c > index a7683d35299..80e897d5f29 100644 > --- a/builtin/submodule--helper.c > +++ b/builtin/submodule--helper.c > @@ -2247,6 +2247,7 @@ static int run_update_command(const struct update_data *ud, int subforce) > case SM_UPDATE_COMMAND: > cp.use_shell = 1; > strvec_push(&cp.args, ud->update_strategy.command); > + BUG("unused??"); > break; > default: > BUG("unexpected update strategy type: %d", > > Will show you that this is used, i.e. this failure in > t7406-submodule-update.sh: > > [...] > + cd super > + git config submodule.submodule.update !git checkout > + cd super/submodule > + git reset --hard HEAD^ > HEAD is now at 0c90624 upstream > + cd super > + cd submodule > + compare_head > + git rev-list --max-count=1 main > + sha_main=0c90624ab7f1aaa301d3bb79f60dcfed1ec4897f > + git rev-list --max-count=1 HEAD > + sha_head=0c90624ab7f1aaa301d3bb79f60dcfed1ec4897f > + test 0c90624ab7f1aaa301d3bb79f60dcfed1ec4897f = 0c90624ab7f1aaa301d3bb79f60dcfed1ec4897f > + git submodule update submodule > BUG: builtin/submodule--helper.c:2250: unused?? > Aborted > > I think what you may be confused by is that this isn't allowed in the > in-tree .gitmodules file, and indeed you can see another BUG() in > determine_submodule_update_strategy() for a panic if we'd ever get as > far as reading "command" from there. > > But that's not what this part of the docs dicsusses, as the paragraph > shown above the removed lines in the context notes. -- James McCandlish
