[PATCH] Don't update index while recursing (was Re: Segfault during merge)

From: Dave O <cxreg@pobox.com>
Date: May 9, 2009, 05:30 UTC
Message-ID: <alpine.DEB.2.00.0905082224450.30999@narbuckle.genericorp.net>
In-Reply-To: <alpine.DEB.2.00.0905081624230.30999@narbuckle.genericorp.net>

On Fri, 8 May 2009, Dave O wrote:

> Once again, I don't really know what the implications of the index
> operations that are happening here are, but the update_stages() call
> in a recursive merge must be doing surprising.

After writing this, I took another look around merge-recursive.c, and realized that all the calls to update_stages() except this one were careful only to do it when o->call_depth was 0. This simple patch seems to fully rectify the problem.

---
  merge-recursive.c |   11 ++++++-----
  1 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/merge-recursive.c b/merge-recursive.c
index a3721ef..f5df9b9 100644
--- a/merge-recursive.c
+++ b/merge-recursive.c
@@ -933,11 +933,12 @@ static int process_renames(struct merge_options *o,
  				       ren1_src, ren1_dst, branch1,
  				       branch2);
  				update_file(o, 0, ren1->pair->two->sha1, ren1->pair->two->mode, ren1_dst);
-				update_stages(ren1_dst, NULL,
-						branch1 == o->branch1 ?
-						ren1->pair->two : NULL,
-						branch1 == o->branch1 ?
-						NULL : ren1->pair->two, 1);
+				if (!o->call_depth)
+					update_stages(ren1_dst, NULL,
+							branch1 == o->branch1 ?
+							ren1->pair->two : NULL,
+							branch1 == o->branch1 ?
+							NULL : ren1->pair->two, 1);
  			} else if (!sha_eq(dst_other.sha1, null_sha1)) {
  				const char *new_path;
  				clean_merge = 0;

-- 
1.6.3.dirty