# [PATCH] log: use true parents for diff when walking reflogs 3 messages from 2013-08-03 to 2013-08-05. Participants: Thomas Rast, Junio C Hamano. Thread: https://gitlist.dev/t/34600 ## Thomas Rast, 2013-08-03 10:36 Subject: [PATCH] log: use true parents for diff when walking reflogs Message-ID: URL: https://gitlist.dev/e/d6dadc4ab54d81490ca46bcfbd44a61be24f6eb7.1375524982.git.trast%40inf.ethz.ch ``` The reflog walking logic (git log -g) replaces the true parent list with the preceding commit in the reflog. This results in bogus commit diffs when combined with options such as -p; the diff is against the reflog predecessor, not the parent of the commit. Save the true parents on the side, extending the functions from the previous commit. The diff logic picks them up and uses them to show the correct diffs. We do have to be somewhat careful about repeated calling of save_parents(), since the reflog may list a commit more than once. We now store (commit_list*)-1 to distinguish the "not saved yet" and "root commit" cases. This lets us preserve an empty parent list even if save_parents() is repeatedly called. Suggested-by: Jeff King Signed-off-by: Thomas Rast --- Jeff King wrote: > > Your description (and solution) make a lot of sense to me. Another code > path that has a similar problem is the "-g" reflog walker. It rewrites > the parents based on the reflog, and the diffs it produces are mostly > useless (e.g., try "git stash list -p"). > > Should we be applying the same technique there? Good point. This is how. It applies on top of the other patch. It doesn't really help for 'git stash list -p', though, because stashes are merge commits. Now they just don't show anything. You could try 'git stash list -p -m', though. revision.c | 28 +++++++++++++++++++++++++--- t/t1411-reflog-show.sh | 22 ++++++++++++++++++++++ 2 files changed, 47 insertions(+), 3 deletions(-) diff --git a/revision.c b/revision.c index e3ca936..ac20d1a 100644 --- a/revision.c +++ b/revision.c @@ -2848,6 +2848,7 @@ static struct commit *get_revision_1(struct rev_info *revs) free(entry); if (revs->reflog_info) { + save_parents(revs, commit); fake_reflog_parent(revs->reflog_info, commit); commit->object.flags &= ~(ADDED | SEEN | SHOWN); } @@ -3083,6 +3084,8 @@ void put_revision_mark(const struct rev_info *revs, const struct commit *commit) define_commit_slab(saved_parents, struct commit_list *); +#define EMPTY_PARENT_LIST ((struct commit_list *)-1) + void save_parents(struct rev_info *revs, struct commit *commit) { struct commit_list **pp; @@ -3093,16 +3096,35 @@ void save_parents(struct rev_info *revs, struct commit *commit) } pp = saved_parents_at(revs->saved_parents_slab, commit); - assert(*pp == NULL); - *pp = copy_commit_list(commit->parents); + + /* + * When walking with reflogs, we may visit the same commit + * several times: once for each appearance in the reflog. + * + * In this case, save_parents() will be called multiple times. + * We want to keep only the first set of parents. We need to + * store a sentinel value for an empty (i.e., NULL) parent + * list to distinguish it from a not-yet-saved list, however. + */ + if (*pp) + return; + if (commit->parents) + *pp = copy_commit_list(commit->parents); + else + *pp = EMPTY_PARENT_LIST; } struct commit_list *get_saved_parents(struct rev_info *revs, const struct commit *commit) { + struct commit_list *parents; + if (!revs->saved_parents_slab) return commit->parents; - return *saved_parents_at(revs->saved_parents_slab, commit); + parents = *saved_parents_at(revs->saved_parents_slab, commit); + if (parents == EMPTY_PARENT_LIST) + return NULL; + return parents; } void free_saved_parents(struct rev_info *revs) diff --git a/t/t1411-reflog-show.sh b/t/t1411-reflog-show.sh index 9a105fe..6f47c0d 100755 --- a/t/t1411-reflog-show.sh +++ b/t/t1411-reflog-show.sh @@ -144,4 +144,26 @@ test_expect_success 'empty reflog file' ' test_cmp expect actual ' +# This guards against the alternative of showing the diffs vs. the +# reflog ancestor. The reflog used is designed to list the commits +# more than once, so as to exercise the corresponding logic. +test_expect_success 'git log -g -p shows diffs vs. parents' ' + test_commit two && + git branch flipflop && + git update-ref refs/heads/flipflop -m flip1 HEAD^ && + git update-ref refs/heads/flipflop -m flop1 HEAD && + git update-ref refs/heads/flipflop -m flip2 HEAD^ && + git log -g -p flipflop >reflog && + grep -v ^Reflog reflog >actual && + git log -1 -p HEAD^ >log.one && + git log -1 -p HEAD >log.two && + ( + cat log.one; echo + cat log.two; echo + cat log.one; echo + cat log.two + ) >expect && + test_cmp expect actual +' + test_done -- 1.8.4.rc1.414.g34bc5b2 ``` ## Thomas Rast, 2013-08-03 10:40 Subject: Re: [PATCH] log: use true parents for diff when walking reflogs Message-ID: <87mwoz6os8.fsf@hexa.v.cablecom.net> URL: https://gitlist.dev/e/87mwoz6os8.fsf%40hexa.v.cablecom.net In-Reply-To: ``` Doh, once again I forgot the in-reply-to. This patch continues the thread after <20130731225520.GB25882@sigill.intra.peff.net>. This is -- strangely enough -- missing from gmane; its immediate predecessor is this: http://thread.gmane.org/gmane.comp.version-control.git/230968/focus=231453 Sorry for the noise. -- Thomas Rast trast@{inf,student}.ethz.ch ``` ## Junio C Hamano, 2013-08-05 19:16 Subject: Re: [PATCH] log: use true parents for diff when walking reflogs Message-ID: <7vd2psj6dq.fsf@alter.siamese.dyndns.org> URL: https://gitlist.dev/e/7vd2psj6dq.fsf%40alter.siamese.dyndns.org In-Reply-To: ``` Thomas Rast writes: > The reflog walking logic (git log -g) replaces the true parent list > with the preceding commit in the reflog. This results in bogus commit > diffs when combined with options such as -p; the diff is against the > reflog predecessor, not the parent of the commit. > > Save the true parents on the side, extending the functions from the > previous commit. The diff logic picks them up and uses them to show > the correct diffs. > > We do have to be somewhat careful about repeated calling of > save_parents(), since the reflog may list a commit more than once. We > now store (commit_list*)-1 to distinguish the "not saved yet" and > "root commit" cases. This lets us preserve an empty parent list even > if save_parents() is repeatedly called. > > Suggested-by: Jeff King > Signed-off-by: Thomas Rast > --- > > Jeff King wrote: >> >> Your description (and solution) make a lot of sense to me. Another code >> path that has a similar problem is the "-g" reflog walker. It rewrites >> the parents based on the reflog, and the diffs it produces are mostly >> useless (e.g., try "git stash list -p"). >> >> Should we be applying the same technique there? > > Good point. This is how. It applies on top of the other patch. Thanks. > It doesn't really help for 'git stash list -p', though, because > stashes are merge commits. Now they just don't show anything. > could try 'git stash list -p -m', though. Using --first-parent may be more convenient and useful. ```