From patchwork Tue Sep 13 16:13:27 2016
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: [6,of,7] revset: fix order of nested 'or' expression (BC)
From: Yuya Nishihara <yuya@tcha.org>
X-Patchwork-Id: 16607
Message-Id: <14af6142b21b0f3e3cf7.1473783207@mimosa>
To: mercurial-devel@mercurial-scm.org
Date: Wed, 14 Sep 2016 01:13:27 +0900

# HG changeset patch
# User Yuya Nishihara <yuya@tcha.org>
# Date 1466932632 -32400
#      Sun Jun 26 18:17:12 2016 +0900
# Node ID 14af6142b21b0f3e3cf742410d4e7adb92ff4c76
# Parent  fbceee41a25ddcbb81c36489b0b6b183c5e9197f
revset: fix order of nested 'or' expression (BC)

This fixes the order of 'x & (y + z)' where 'y' and 'z' are not trivial.

The follow-order 'or' operation is slower than the ordered operation if
an input set is large:

       #0           #1           #2           #3
    0) 0.002968     0.002980     0.002982     0.073042
    1) 0.004513     0.004485     0.012029     0.075261

    #0: 0:4000 & (0:1099 + 1000:2099 + 2000:3099)
    #1: 4000:0 & (0:1099 + 1000:2099 + 2000:3099)
    #2: 10000:0 & (0:1099 + 1000:2099 + 2000:3099)
    #3: file("path:hg") & (0:1099 + 1000:2099 + 2000:3099)

I've tried another implementation, but which appeared to be slower than
this version.

    ss = [getset(repo, fullreposet(repo), x) for x in xs]
    return subset.filter(lambda r: any(r in s for s in ss), cache=False)

diff --git a/mercurial/revset.py b/mercurial/revset.py
--- a/mercurial/revset.py
+++ b/mercurial/revset.py
@@ -407,7 +407,12 @@ def _orsetlist(repo, subset, xs):
     return a + b
 
 def orset(repo, subset, x, order):
-    return _orsetlist(repo, subset, getlist(x))
+    xs = getlist(x)
+    if order == followorder:
+        # slow path to take the subset order
+        return subset & _orsetlist(repo, fullreposet(repo), xs)
+    else:
+        return _orsetlist(repo, subset, xs)
 
 def notset(repo, subset, x, order):
     return subset - getset(repo, subset, x)
diff --git a/tests/test-revset.t b/tests/test-revset.t
--- a/tests/test-revset.t
+++ b/tests/test-revset.t
@@ -1317,15 +1317,14 @@ ordering defined by it.
       follow)
     define)
   * set:
-  <addset
-    <filteredset
+  <filteredset
+    <spanset- 0:2>,
+    <addset
       <spanset+ 0:1>,
-      <spanset- 0:2>>,
-    <baseset [2]>>
+      <baseset [2]>>>
+  2
+  1
   0
-  1
-  2
- BROKEN: should be '2 1 0'
 
  '_intlist(a b)' should behave like 'a + b':