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.
#!/bin/sh
hg init
echo a > a
hg commit -A -ma
echo a >> a
hg commit -mb
echo a >> a
hg commit -mc
hg up 1
echo a >> a
hg commit -md
hg up 1
echo a >> a
hg commit -me
hg up 1
echo % should fail because not at a head
hg merge
hg up
echo % should fail because \> 2 heads
hg merge
echo % should succeed
hg merge 2
hg commit -mm1
echo % should succeed - 2 heads
hg merge
hg commit -mm2
echo % should fail because at tip
hg merge
hg up 0
echo % should fail because 1 head
hg merge
true