On Wed, Apr 12, 2023 at 12:22:44PM +0200, Patrick Steinhardt wrote: > When passed the same packfile twice via `--stdin-packs` we return an > error that the packfile supposedly was not found. This is because when > reading packs into the list of included or excluded packfiles, we will > happily re-add packfiles even if they are part of the lists already. And > while the list can now contain duplicates, we will only set the `util` > pointer of the first list entry to the `packed_git` structure. We notice > that at a later point when checking that all list entries have their > `util` pointer set and die with an error. Well explained. > Signed-off-by: Patrick Steinhardt <ps@xxxxxx> > --- > builtin/pack-objects.c | 2 ++ > t/t5331-pack-objects-stdin.sh | 29 +++++++++++++++++++++++++++++ > t/t7703-repack-geometric.sh | 25 +++++++++++++++++++++++++ > 3 files changed, 56 insertions(+) > create mode 100755 t/t5331-pack-objects-stdin.sh > > diff --git a/builtin/pack-objects.c b/builtin/pack-objects.c > index 77d88f85b0..fdf3f440be 100644 > --- a/builtin/pack-objects.c > +++ b/builtin/pack-objects.c > @@ -3359,7 +3359,9 @@ static void read_packs_list_from_stdin(void) > } > > string_list_sort(&include_packs); > + string_list_remove_duplicates(&include_packs, 0); > string_list_sort(&exclude_packs); > + string_list_remove_duplicates(&exclude_packs, 0); I for some reason thought that string_list_remove_duplicates() sorted its first argument as a side-effect, but double checking on it, that isn't the case. So I agree that calling string_list_remove_duplicates() is the right thing to do here. > diff --git a/t/t5331-pack-objects-stdin.sh b/t/t5331-pack-objects-stdin.sh > new file mode 100755 > index 0000000000..ab34cfc729 > --- /dev/null > +++ b/t/t5331-pack-objects-stdin.sh Cool, I am glad that we are moving some of these tests out of t5300. We could probably go a little bit further, since there are a handful of purely `--stdin-packs`-related tests towards the bottom of that script that could or should also be moved out. We could also always do it later, but I wouldn't be sad to see an extra commit before this one that introduces t5331 and moves the last six or so `--stdin-packs` tests out of t5300. If nothing else, it would be nice to be able to run the tests for just that feature without having to skip a bunch of other unrelated tests. Luckily, those tests don't depend on any of the earlier ones or setup tests, so they should be able to move with a pure cut and paste. > @@ -0,0 +1,29 @@ > +#!/bin/sh > + > +test_description='pack-objects --stdin' > +GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME=main > +export GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME > + > +TEST_PASSES_SANITIZE_LEAK=true > +. ./test-lib.sh > + > +test_expect_success 'pack-objects --stdin with duplicate packfile' ' > + test_when_finished "rm -fr repo" && > + > + git init repo && > + ( > + cd repo && > + test_commit "commit" && > + git repack -ad && > + > + ( > + basename .git/objects/pack/pack-*.pack && > + basename .git/objects/pack/pack-*.pack > + ) >packfiles && > + > + git pack-objects --stdin-packs generated-pack <packfiles && > + test_cmp generated-pack-*.pack .git/objects/pack/pack-*.pack I think that in practice that test_cmp'ing these two together will be just fine, but if you wanted to be extra careful, you could grab their object lists and compare those two together. That would allow you to continue passing this test without needing to worry about if you generated the same logical content with slightly different on-disk content. Thanks, Taylor
