git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [BUG] git-log shows first parent and repeated last for octopus merge
@ 2006-10-18 11:11 Jakub Narebski
  2006-10-18 13:53 ` Johannes Schindelin
  2006-10-18 16:35 ` A Large Angry SCM
  0 siblings, 2 replies; 6+ messages in thread
From: Jakub Narebski @ 2006-10-18 11:11 UTC (permalink / raw)
  To: git

When trying to find how many merges and how many octopus merges (merges with
more than two parents) are in git.git repository I have encountered the
following strange output of git-log:

 1000:jnareb@roke:~/git> git log --parents --full-history --max-count=1 \
   211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
 commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e <last parent repeated>
 Merge: d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b...
 [...]

while git-show gives correct output

 1005:jnareb@roke:~/git> git show 211232bae64bcc60bbf5d1b5e5b2344c22ed767e
 commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e
 Merge: fc54a9c... 9e30dd7... c4b83e6... 6602659... b28858b...
 [...]

Same with git-rev-list, also shows correct output:

 1005:jnareb@roke:~/git> git rev-list --header --parents --max-count=1 \
   211232bae64bcc60bbf5d1b5e5b2344c22ed767e
 211232bae64bcc60bbf5d1b5e5b2344c22ed767e [...] 
 tree cdafa88fa4ed7fcc7bb6c64d62e2d7c4d3b65e42
 parent fc54a9c30ccad3fde5890d2c0ca2e2acc0848fbc
 parent 9e30dd7c0ecc9f10372f31539d0122db97418353
 parent c4b83e618f1df7d8ecc9392fa40e5bebccbe6b5a
 parent 660265909fc178581ef327076716dfd3550e6e7b
 parent b28858bf65d4fd6d8bb070865518ec43817fe7f3
 [...]

1008:jnareb@roke:~/git> git --version
git version 1.4.2.1
-- 
Jakub Narebski
Warsaw, Poland
ShadeHawk on #git

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [BUG] git-log shows first parent and repeated last for octopus merge
  2006-10-18 11:11 [BUG] git-log shows first parent and repeated last for octopus merge Jakub Narebski
@ 2006-10-18 13:53 ` Johannes Schindelin
  2006-10-18 14:02   ` Jakub Narebski
  2006-10-18 19:02   ` [BUG] git-log shows first parent and repeated last for octopus merge Junio C Hamano
  2006-10-18 16:35 ` A Large Angry SCM
  1 sibling, 2 replies; 6+ messages in thread
From: Johannes Schindelin @ 2006-10-18 13:53 UTC (permalink / raw)
  To: Jakub Narebski; +Cc: git

Hi,

On Wed, 18 Oct 2006, Jakub Narebski wrote:

> When trying to find how many merges and how many octopus merges (merges with
> more than two parents) are in git.git repository I have encountered the
> following strange output of git-log:
> 
>  1000:jnareb@roke:~/git> git log --parents --full-history --max-count=1 \
>    211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
>  commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e <last parent repeated>
>  Merge: d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b...
>  [...]

This happens because a//b rewrites the history, i.e. the parents are 
edited. IMHO it makes no sense at all to show the parents in such a case, 
since they are bogus.

Ciao,
Dscho

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [BUG] git-log shows first parent and repeated last for octopus merge
  2006-10-18 13:53 ` Johannes Schindelin
@ 2006-10-18 14:02   ` Jakub Narebski
  2006-10-18 14:26     ` [PATCH] When rewriting parents, cull duplicates Johannes Schindelin
  2006-10-18 19:02   ` [BUG] git-log shows first parent and repeated last for octopus merge Junio C Hamano
  1 sibling, 1 reply; 6+ messages in thread
From: Jakub Narebski @ 2006-10-18 14:02 UTC (permalink / raw)
  To: Johannes Schindelin; +Cc: git

Johannes Schindelin wrote:
> On Wed, 18 Oct 2006, Jakub Narebski wrote:
> 
> > When trying to find how many merges and how many octopus merges (merges with
> > more than two parents) are in git.git repository I have encountered the
> > following strange output of git-log:
> > 
> >  1000:jnareb@roke:~/git> git log --parents --full-history --max-count=1 \
> >    211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
> >  commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e <last parent repeated>
> >  Merge: d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b...
> >  [...]
> 
> This happens because a//b rewrites the history, i.e. the parents are 
> edited. IMHO it makes no sense at all to show the parents in such a case, 
> since they are bogus.

Or rather it has no sense to _repeat_ rewritten parent the number of times
the commit has parents originally. Compare git-log and git-rev-list results:

1010:jnareb@roke:~/git> git rev-list --full-history --header --parents \
  --max-count=2 211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
211232bae64bcc60bbf5d1b5e5b2344c22ed767e d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f
tree cdafa88fa4ed7fcc7bb6c64d62e2d7c4d3b65e42
parent fc54a9c30ccad3fde5890d2c0ca2e2acc0848fbc
parent 9e30dd7c0ecc9f10372f31539d0122db97418353
parent c4b83e618f1df7d8ecc9392fa40e5bebccbe6b5a
parent 660265909fc178581ef327076716dfd3550e6e7b
parent b28858bf65d4fd6d8bb070865518ec43817fe7f3
author Junio C Hamano <junkio@cox.net> 1115335014 -0700
committer Junio C Hamano <junkio@cox.net> 1115335014 -0700

    Octopus merge of the following five patches.
    
      Update git-apply-patch-script for symbolic links.
      Make git-prune-script executable again.
      Do not write out new index if nothing has changed.
      diff-cache shows differences for unmerged paths without --cache.
      Update diff engine for symlinks stored in the cache.
    
    Signed-off-by: Junio C Hamano <junkio@cox.net>
d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f 54c26fb9d0cdff94c7717125d0a222b324bfea8a
[...]

1014:jnareb@roke:~/git> PAGER= git log --full-history --parents \
  --max-count=2 211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f
Merge: d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b...
Author: Junio C Hamano <junkio@cox.net>
Date:   Thu May 5 16:16:54 2005 -0700

    Octopus merge of the following five patches.
    
      Update git-apply-patch-script for symbolic links.
      Make git-prune-script executable again.
      Do not write out new index if nothing has changed.
      diff-cache shows differences for unmerged paths without --cache.
      Update diff engine for symlinks stored in the cache.
    
    Signed-off-by: Junio C Hamano <junkio@cox.net>

commit d0d0d0bd3c2c4591ffbc292d7e082e8ad8f2057f 54c26fb9d0cdff94c7717125d0a222b324bfea8a 54c26fb9d0cdff94c7717125d0a222b324bfea8a
[...]

-- 
Jakub Narebski
Poland

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH] When rewriting parents, cull duplicates
  2006-10-18 14:02   ` Jakub Narebski
@ 2006-10-18 14:26     ` Johannes Schindelin
  0 siblings, 0 replies; 6+ messages in thread
From: Johannes Schindelin @ 2006-10-18 14:26 UTC (permalink / raw)
  To: Jakub Narebski; +Cc: git


Earlier, when calling

	git log --parents some-ref -- path/file

it was possible that in case of merges, multiple parents would be
rewritten as the _same_ commit, which would happily be printed
multiple times.

Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>

---

	Subject: Re: [BUG] git-log shows first parent and repeated last 
	for octopus merge

	On Wed, 18 Oct 2006, Jakub Narebski wrote:

	> Johannes Schindelin wrote:
	> > On Wed, 18 Oct 2006, Jakub Narebski wrote:
	> > 
	> > > When trying to find how many merges and how many octopus 
	> > > merges (merges with more than two parents) are in git.git 
	> > > repository I have encountered the following strange output 
	> > > of git-log:
	> > > 
	> > > 1000:jnareb@roke:~/git> git log --parents --full-history \
	> > > --max-count=1 211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
	> > >  commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e <...>
	> > >  Merge: d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b...
	> > >  [...]
	> > 
	> > This happens because a//b rewrites the history, i.e. the 
	> > parents are edited. IMHO it makes no sense at all to show the 
	> > parents in such a case, since they are bogus.
	> 
	> Or rather it has no sense to _repeat_ rewritten parent the 
	> number of times the commit has parents originally.

	Here you are.

 revision.c |   13 ++++++++++++-
 1 files changed, 12 insertions(+), 1 deletions(-)

diff --git a/revision.c b/revision.c
index 280e92b..8a2ca52 100644
--- a/revision.c
+++ b/revision.c
@@ -1086,12 +1086,23 @@ static int rewrite_one(struct rev_info *
 	}
 }
 
+static int parent_is_duplicate(struct commit_list *parents,
+		struct commit_list *current)
+{
+	for (; parents != current; parents = parents->next)
+		if (parents->item == current->item)
+			return 1;
+	return 0;
+}
+
 static void rewrite_parents(struct rev_info *revs, struct commit *commit)
 {
 	struct commit_list **pp = &commit->parents;
 	while (*pp) {
 		struct commit_list *parent = *pp;
-		if (rewrite_one(revs, &parent->item) < 0) {
+		if (rewrite_one(revs, &parent->item) < 0 ||
+				/* cull duplicates */
+				parent_is_duplicate(commit->parents, parent)) {
 			*pp = parent->next;
 			continue;
 		}
-- 
1.4.2.4.g21cef-dirty

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [BUG] git-log shows first parent and repeated last for octopus merge
  2006-10-18 11:11 [BUG] git-log shows first parent and repeated last for octopus merge Jakub Narebski
  2006-10-18 13:53 ` Johannes Schindelin
@ 2006-10-18 16:35 ` A Large Angry SCM
  1 sibling, 0 replies; 6+ messages in thread
From: A Large Angry SCM @ 2006-10-18 16:35 UTC (permalink / raw)
  To: Jakub Narebski; +Cc: git

Jakub Narebski wrote:
> When trying to find how many merges and how many octopus merges (merges with
> more than two parents) are in git.git repository I have encountered the
> following strange output of git-log:

Maybe the following will find what you're looking for:

git-rev-list master --parents \
	|sed -e 's/[0-9a-f]\{40\}//' -e 's/ [0-9a-f]\{40\}/*/g' \
	|sort \
	|uniq -c

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [BUG] git-log shows first parent and repeated last for octopus merge
  2006-10-18 13:53 ` Johannes Schindelin
  2006-10-18 14:02   ` Jakub Narebski
@ 2006-10-18 19:02   ` Junio C Hamano
  1 sibling, 0 replies; 6+ messages in thread
From: Junio C Hamano @ 2006-10-18 19:02 UTC (permalink / raw)
  To: Johannes Schindelin; +Cc: git

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

> Hi,
>
> On Wed, 18 Oct 2006, Jakub Narebski wrote:
>
>> When trying to find how many merges and how many octopus merges (merges with
>> more than two parents) are in git.git repository I have encountered the
>> following strange output of git-log:
>> 
>>  1000:jnareb@roke:~/git> git log --parents --full-history --max-count=1 \
>>    211232bae64bcc60bbf5d1b5e5b2344c22ed767e -- a//b
>>  commit 211232bae64bcc60bbf5d1b5e5b2344c22ed767e <last parent repeated>
>>  Merge: d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b... d0d0d0b...
>>  [...]
>
> This happens because a//b rewrites the history, i.e. the parents are 
> edited. IMHO it makes no sense at all to show the parents in such a case, 
> since they are bogus.

If the command line did not have --full-history, I would agree
with you.  The caller explicitly told us not to remove side
branches that do not end up modifying the paths, and also told
us, with --parent, to show parenthood information after removing
intermediate commits that do not change the tree shape (with
respect to the specified paths).

So it is showing what the user asked -- the user may not have
understood what he was asking, but that is a separate problem, I
would think.

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2006-10-18 19:02 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-10-18 11:11 [BUG] git-log shows first parent and repeated last for octopus merge Jakub Narebski
2006-10-18 13:53 ` Johannes Schindelin
2006-10-18 14:02   ` Jakub Narebski
2006-10-18 14:26     ` [PATCH] When rewriting parents, cull duplicates Johannes Schindelin
2006-10-18 19:02   ` [BUG] git-log shows first parent and repeated last for octopus merge Junio C Hamano
2006-10-18 16:35 ` A Large Angry SCM

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).