On 29. juni 2010, at 18.19, Junio C Hamano wrote:

quoted text
> I do however wonder if this should be initially introduced as an
> experimental feature, guarded with a configuration option for brave souls
> to try it out, and flip the feature on by default after we gain confidence
> in it, both in performance and in correctness.

Agreed, and thanks.  Messing with the merge machinery unnerves me a little.

Shouldn't the normalization in merge-recursive be conditional too?

Something like this squashed into the delete/modify patch:

--8<--
diff --git a/merge-recursive.c b/merge-recursive.c
index a2c174f..49bd3d2 100644
--- a/merge-recursive.c
+++ b/merge-recursive.c
@@ -1079,6 +1079,8 @@ static int normalized_eq(const unsigned char *a_sha,
 {
        struct strbuf a = STRBUF_INIT;
        struct strbuf b = STRBUF_INIT;
+       if (!core_ll_merge_double_convert)
+               return 0;
        int ret = 0;
        if (a_sha && b_sha &&
            read_sha1_strbuf(a_sha, path, &a) &&
--8<--

Sorry about the whitespace-damaged inline diff, I'll redo the series if necessary.
-- 
Eyvind

--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html