setdiscovery: always add exponential sample to the heads
As explained in a previous changeset, prioritizing heads too much behaves
pathologically when there are more heads than the sample size. To counter this,
we always inject exponential samples before reducing to the sample size limit.
This already show some benefit in the test themselves, but on a real-world example
this moves my discovery for push to pathologically headed repo from 45 rounds to
17 of them.
We should maybe ensure that at least 25% of the result sample is heads, but I
think the random sampling will be fine in practice.
--- a/mercurial/setdiscovery.py Wed Jan 07 17:23:21 2015 -0800
+++ b/mercurial/setdiscovery.py Wed Jan 07 17:28:51 2015 -0800
@@ -105,10 +105,7 @@
return sample
def _takefullsample(dag, nodes, size):
- always = dag.headsetofconnecteds(nodes)
- if size <= len(always):
- return always
- sample = always
+ sample = always = dag.headsetofconnecteds(nodes)
# update from heads
_updatesample(dag, nodes, sample, always)
# update from roots
--- a/tests/test-setdiscovery.t Wed Jan 07 17:23:21 2015 -0800
+++ b/tests/test-setdiscovery.t Wed Jan 07 17:28:51 2015 -0800
@@ -317,17 +317,14 @@
query 3; still undecided: 1140, sample size is: 200
sampling from both directions
searching: 4 queries
- query 4; still undecided: 940, sample size is: 200
+ query 4; still undecided: 592, sample size is: 200
sampling from both directions
searching: 5 queries
- query 5; still undecided: 740, sample size is: 200
+ query 5; still undecided: 292, sample size is: 200
sampling from both directions
searching: 6 queries
- query 6; still undecided: 540, sample size is: 200
- sampling from both directions
- searching: 7 queries
- query 7; still undecided: 46, sample size is: 46
- 7 total queries
+ query 6; still undecided: 51, sample size is: 51
+ 6 total queries
common heads: 3ee37d65064a
Test actual protocol when pulling one new head in addition to common heads