| Message ID | 20241206-pks-rev-parse-fix-reversed-list-v3-1-d934c17db168@pks.im (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 0ff919e87a08b7ab81507917ca55eb613296d043 |
| Headers | show |
| Series | [v3] object-name: fix reversed ordering with ":/<PATTERN>" revisions | expand |
Just bringing up this nitpick in case you decide to make a v4 based on Justin’s reply in this thread. On Fri, Dec 6, 2024, at 16:45, Patrick Steinhardt wrote: > Recently it was reported [1] that "look for the youngest reachable > commit with log message that match the given pattern" syntax (e.g. > ':/<PATTERN>') started to return results in reverse recency order. This This isn’t “e.g. ':/<PATTERN>'” any more since you are not listing examples. “i.e.” perhaps? But it’s clear what you mean from the context anyway.
"Kristoffer Haugsbakk" <kristofferhaugsbakk@fastmail.com> writes: > This isn’t “e.g. ':/<PATTERN>'” any more since you are not listing > examples. “i.e.” perhaps? Will locally amend (and with s/PATTERN/text/ to match the relevant part of the documentation). Thanks for carefully reading.
Am 06.12.24 um 16:45 schrieb Patrick Steinhardt: > @@ -1423,7 +1423,7 @@ static int get_oid_oneline(struct repository *r, > > for (l = list; l; l = l->next) { > l->item->object.flags |= ONELINE_SEEN; > - commit_list_insert(l->item, ©); > + copy_tail = &commit_list_insert(l->item, copy_tail)->next; OK. The following does the same while being clearer: copy_tail = commit_list_append(l->item, copy_tail); You could get the idea to do that replacement across the whole source tree. That would be nice, but must not be done blindly (e.g. with Coccinelle), as the result will be different if the second argument can point to somewhere in the middle of the list. Here it's OK because we indeed are appending (invariant *copy_tail == NULL holds). René
diff --git a/object-name.c b/object-name.c index c892fbe80aa7173dfcc1995de5a75bc322c6adb7..34433d2a01d6a23ce6b4ca19b85c53b7b82fd0e5 100644 --- a/object-name.c +++ b/object-name.c @@ -1401,7 +1401,7 @@ static int get_oid_oneline(struct repository *r, const char *prefix, struct object_id *oid, const struct commit_list *list) { - struct commit_list *copy = NULL; + struct commit_list *copy = NULL, **copy_tail = © const struct commit_list *l; int found = 0; int negative = 0; @@ -1423,7 +1423,7 @@ static int get_oid_oneline(struct repository *r, for (l = list; l; l = l->next) { l->item->object.flags |= ONELINE_SEEN; - commit_list_insert(l->item, ©); + copy_tail = &commit_list_insert(l->item, copy_tail)->next; } while (copy) { const char *p, *buf; diff --git a/t/t1500-rev-parse.sh b/t/t1500-rev-parse.sh index 30c31918fde6539d52800e18dfbb3423b5b73491..42c4a63cb95eed781ed7d3029c4ff5e600e6f8b8 100755 --- a/t/t1500-rev-parse.sh +++ b/t/t1500-rev-parse.sh @@ -310,4 +310,19 @@ test_expect_success '--short= truncates to the actual hash length' ' test_cmp expect actual ' +test_expect_success ':/ and HEAD^{/} favor more recent matching commits' ' + test_when_finished "rm -rf repo" && + git init repo && + ( + cd repo && + test_commit common-old && + test_commit --no-tag common-new && + git rev-parse HEAD >expect && + git rev-parse :/common >actual && + test_cmp expect actual && + git rev-parse HEAD^{/common} >actual && + test_cmp expect actual + ) +' + test_done
Recently it was reported [1] that "look for the youngest reachable commit with log message that match the given pattern" syntax (e.g. ':/<PATTERN>') started to return results in reverse recency order. This regression was introduced in Git v2.47.0 and is caused by a memory leak fix done in 57fb139b5e (object-name: fix leaking commit list items, 2024-08-01). The intent of the identified commit is to stop modifying the commit list provided by the caller such that the caller can properly free all commit list items, including those that the called function might potentially remove from the list. This was done by creating a copy of the passed-in commit list and modifying this copy instead of the caller-provided list. We already knew to create such a copy beforehand with the `backup` list, which was used to clear the `ONELINE_SEEN` commit mark after we were done. So the refactoring simply renamed that list to `copy` and started to operate on that list instead. There is a gotcha though: the backup list, and thus now also the copied list, is always being prepended to, so the resulting list is in reverse order! The end result is that we pop commits from the wrong end of the commit list, returning commits in reverse recency order. Fix the bug by appending to the list instead. [1]: <CAKOEJdcPYn3O01p29rVa+xv=Qr504FQyKJeSB-Moze04ViCGGg@mail.gmail.com> Reported-by: Aarni Koskela <aarni@valohai.com> Signed-off-by: Patrick Steinhardt <ps@pks.im> --- This patch applies on top of v2.47.0, which is the first version which had this regression. Changes in v2: - Include the message ID of the report in the commit message. - Fix terminology used by the commit message. - Move the test from t4208 to t1500. - Link to v1: https://lore.kernel.org/r/20241206-pks-rev-parse-fix-reversed-list-v1-1-95a96564a4d7@pks.im Changes in v3: - Only mention ':/' as having regressed, not 'HEAD^{/}'. - Link to v2: https://lore.kernel.org/r/20241206-pks-rev-parse-fix-reversed-list-v2-1-190514278ead@pks.im Thanks! Patrick --- object-name.c | 4 ++-- t/t1500-rev-parse.sh | 15 +++++++++++++++ 2 files changed, 17 insertions(+), 2 deletions(-) --- base-commit: 777489f9e09c8d0dd6b12f9d90de6376330577a2 change-id: 20241206-pks-rev-parse-fix-reversed-list-0f94a20a6165