I have spotted the biggest bottleneck in "bdiff.c". Actually it was
pretty easy to find after I recompiled the python interpreter and
mercurial for profiling.
In "bdiff.c" function "equatelines" allocates the minimum hash table
size, which can lead to tons of collisions. I introduced an
"overcommit" factor of 16, this is, I allocate 16 times more memory
than the minimum value. Overcommiting 128 times does not improve the
performance over the 16-times case.
quiet:
--- a/foo Mon Jan 12 13:46:40 1970 +0000
+++ b/foo Mon Jan 12 13:46:41 1970 +0000
@@ -1,1 +1,1 @@ bar
-bar
+foobar
normal:
diff -r 74de3f1392e2 -r b8b5f023a6ad foo
--- a/foo Mon Jan 12 13:46:40 1970 +0000
+++ b/foo Mon Jan 12 13:46:41 1970 +0000
@@ -1,1 +1,1 @@ bar
-bar
+foobar
verbose:
diff -r 74de3f1392e2 -r b8b5f023a6ad foo
--- a/foo Mon Jan 12 13:46:40 1970 +0000
+++ b/foo Mon Jan 12 13:46:41 1970 +0000
@@ -1,1 +1,1 @@ bar
-bar
+foobar
debug:
diff -r 74de3f1392e2d67856fb155963441f2610494e1a -r b8b5f023a6ad77fc378bd95cf3fa00cd1414d107 foo
--- a/foo Mon Jan 12 13:46:40 1970 +0000
+++ b/foo Mon Jan 12 13:46:41 1970 +0000
@@ -1,1 +1,1 @@ bar
-bar
+foobar