Consider two commits on different branches, one with this patch:
diff --git a/file.txt b/file.txt
index 704fa27..2f7e74c 100644
--- a/file.txt
+++ b/file.txt
@@ -1,3 +1,3 @@
old_context
-foo
+bar
and the other with this patch:
diff --git a/file.txt b/file.txt
index f35051b..8c7de32 100644
--- a/file.txt
+++ b/file.txt
@@ -1,3 +1,3 @@
new_context
-foo
+bar
If I run "git rev-list --cherry-pick --left-right branch1...branch2", it
reports both commits as being genuine commits on their respective
branch, even though I consider their patches to be the same.
I guess for my purpose I would like to have patch-ids that ignore
context (or that use only one line of context, I'm not sure which).
In fact, if I do "git show <commit> -U1 | git patch-id", both commits
show the same id.
So, would it make sense to have a parameter for git-rev-list (and
git-cherry) that lets you specify how much context to be used for the
patch ids?
--
Stefan Haller
Berlin, Germany
http://www.haller-berlin.de/
Stefan Haller venit, vidit, dixit 02.09.2011 12:35:
Consider two commits on different branches, one with this patch:
diff --git a/file.txt b/file.txt
index 704fa27..2f7e74c 100644
--- a/file.txt
+++ b/file.txt
@@ -1,3 +1,3 @@
old_context
-foo
+bar
and the other with this patch:
diff --git a/file.txt b/file.txt
index f35051b..8c7de32 100644
--- a/file.txt
+++ b/file.txt
@@ -1,3 +1,3 @@
new_context
-foo
+bar
If I run "git rev-list --cherry-pick --left-right branch1...branch2", it
reports both commits as being genuine commits on their respective
branch, even though I consider their patches to be the same.
I guess for my purpose I would like to have patch-ids that ignore
context (or that use only one line of context, I'm not sure which).
In fact, if I do "git show <commit> -U1 | git patch-id", both commits
show the same id.
So, would it make sense to have a parameter for git-rev-list (and
git-cherry) that lets you specify how much context to be used for the
patch ids?
It would be a bit like the patch below. "git log" accepts diff options already.
But:
- Do we want the patch id generation and the patch display (-p) to use the
same options?
- -U1 implies -p/--patch and there is no --no-patch.
- Which other diff options do we want to pass to the patch id
generation: --histogram, --patience, ...?
Cheers,
Michael
----
diff --git i/diff.c w/diff.c
index fcc0078..4e82912 100644
--- i/diff.c
+++ w/diff.c
@@ -4103,7 +4103,7 @@ static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
}
xpp.flags = 0;
- xecfg.ctxlen = 3;
+ xecfg.ctxlen = options->context;
xecfg.flags = 0;
xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
&xpp, &xecfg);
diff --git i/revision.c w/revision.c
index 072ddac..5a98ed9 100644
--- i/revision.c
+++ w/revision.c
@@ -601,6 +601,7 @@ static void cherry_pick_list(struct commit_list *list, struct rev_info *revs)
left_first = left_count < right_count;
init_patch_ids(&ids);
ids.diffopts.pathspec = revs->diffopt.pathspec;
+ ids.diffopts.context = revs->diffopt.context;
/* Compute patch-ids for one side */
for (p = list; p; p = p->next) {Michael J Gruber [off-list ref] wrote:
Stefan Haller venit, vidit, dixit 02.09.2011 12:35:
quoted
Consider two commits on different branches, one with this patch:
diff --git a/file.txt b/file.txt
index 704fa27..2f7e74c 100644
--- a/file.txt
+++ b/file.txt
@@ -1,3 +1,3 @@
old_context
-foo
+bar
and the other with this patch:
diff --git a/file.txt b/file.txt
index f35051b..8c7de32 100644
--- a/file.txt
+++ b/file.txt
@@ -1,3 +1,3 @@
new_context
-foo
+bar
If I run "git rev-list --cherry-pick --left-right branch1...branch2", it
reports both commits as being genuine commits on their respective
branch, even though I consider their patches to be the same.
I guess for my purpose I would like to have patch-ids that ignore
context (or that use only one line of context, I'm not sure which).
In fact, if I do "git show <commit> -U1 | git patch-id", both commits
show the same id.
So, would it make sense to have a parameter for git-rev-list (and
git-cherry) that lets you specify how much context to be used for the
patch ids?
It would be a bit like the patch below. "git log" accepts diff options already.
But:
[...]
Thanks a lot. I can't contribute much to answering your "But:"
questions; I can only add more questions myself. :-)
Is there a reason why the hard-coded default is 3 in the current code?
It seems to me that 1 would be a better choice; it would mean "patches
are equal if their added/removed lines are the same, and they could be
cherry-picked without conflicts."
Now, I'm in a situation where I'll be stuck with git 1.7.1 for quite a
while, so no patch is going to help me. It looks like the only way to
get the behaviour I want is to reimplement git-rev-list --cherry-pick
myself, feeding each patch to git-patch-id, right? (Horribly
inefficient, but might be good enough for my purpose. Just wondering if
I'm missing a smarter way to solve it.)
Thanks,
Stefan
--
Stefan Haller
Berlin, Germany
http://www.haller-berlin.de/