changeset 23343:f8a2647fe020

setdiscovery: avoid a full changelog graph traversal We were definitely being suboptimal here: we were constructing two full sets, one with the full set of common nodes (i.e. a graph traversal) and one with all nodes. Then we subtract one set from the other. This whole process is O(commits) and causes discovery to be significantly slower than it should be. Instead, keep track of common incrementally and keep undecided as small as possible. This makes discovery massively faster on large repos: on one such repo, 'hg debugdiscovery' over SSH with one commit missing on the client and five on the server went from 4.5 seconds to 1.5. (An 'hg debugdiscovery' with no commits missing on the client, i.e. connection startup time, was 1.2 seconds.)
author Siddharth Agarwal <sid0@fb.com>
date Sun, 16 Nov 2014 00:40:29 -0800
parents f710644e1ce9
children 49cdf51cbc6c
files mercurial/setdiscovery.py
diffstat 1 files changed, 19 insertions(+), 13 deletions(-) [+]
line wrap: on
line diff
--- a/mercurial/setdiscovery.py	Fri Nov 14 19:40:30 2014 -0800
+++ b/mercurial/setdiscovery.py	Sun Nov 16 00:40:29 2014 -0800
@@ -40,7 +40,7 @@
 classified with it (since all ancestors or descendants will be marked as well).
 """
 
-from node import nullid
+from node import nullid, nullrev
 from i18n import _
 import random
 import util, dagutil
@@ -177,27 +177,23 @@
     # own nodes where I don't know if remote knows them
     undecided = dag.nodeset()
     # own nodes I know we both know
-    common = set()
+    # treat remote heads (and maybe own heads) as a first implicit sample
+    # response
+    common = cl.incrementalmissingrevs(srvheads)
+    commoninsample = set(n for i, n in enumerate(sample) if yesno[i])
+    common.addbases(commoninsample)
+    undecided = set(common.missingancestors(ownheads))
     # own nodes I know remote lacks
     missing = set()
 
-    # treat remote heads (and maybe own heads) as a first implicit sample
-    # response
-    common.update(dag.ancestorset(srvheads))
-    undecided.difference_update(common)
-
     full = False
     while undecided:
 
         if sample:
-            commoninsample = set(n for i, n in enumerate(sample) if yesno[i])
-            common.update(dag.ancestorset(commoninsample, common))
-
             missinginsample = [n for i, n in enumerate(sample) if not yesno[i]]
             missing.update(dag.descendantset(missinginsample, missing))
 
             undecided.difference_update(missing)
-            undecided.difference_update(common)
 
         if not undecided:
             break
@@ -206,7 +202,7 @@
             ui.note(_("sampling from both directions\n"))
             sample = _takefullsample(dag, undecided, size=fullsamplesize)
             targetsize = fullsamplesize
-        elif common:
+        elif common.hasbases():
             # use cheapish initial sample
             ui.debug("taking initial sample\n")
             sample = _takefullsample(dag, undecided, size=fullsamplesize)
@@ -228,7 +224,17 @@
         yesno = remote.known(dag.externalizeall(sample))
         full = True
 
-    result = dag.headsetofconnecteds(common)
+        if sample:
+            commoninsample = set(n for i, n in enumerate(sample) if yesno[i])
+            common.addbases(commoninsample)
+            common.removeancestorsfrom(undecided)
+
+    # heads(common) == heads(common.bases) since common represents common.bases
+    # and all its ancestors
+    result = dag.headsetofconnecteds(common.bases)
+    # common.bases can include nullrev, but our contract requires us to not
+    # return any heads in that case, so discard that
+    result.discard(nullrev)
     ui.progress(_('searching'), None)
     ui.debug("%d total queries\n" % roundtrips)