Patchwork D1616: rebase: disable `inmemory` if the rebaseset contains the working copy

login
register
mail settings
Submitter phabricator
Date Dec. 8, 2017, 7:25 a.m.
Message ID <differential-rev-PHID-DREV-r66eecchis7zqj2rxsmo-req@phab.mercurial-scm.org>
Download mbox | patch
Permalink /patch/26079/
State Superseded
Headers show

Comments

phabricator - Dec. 8, 2017, 7:25 a.m.
phillco created this revision.
Herald added a subscriber: mercurial-devel.
Herald added a reviewer: hg-reviewers.

REVISION SUMMARY
  As described in the comment, rebasing the working copy parent with in-memory
  merge, and then updating to the new commit, isn't much faster because of the
  extra overhead of uppdating. Best to leavew it off in that case.
  
  This commit makes deploying in-memory merge via an extension easier, because
  you can just set `inmemory=True` based on some config or probability, and this
  will turn off the cases where it's not desired.

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D1616

AFFECTED FILES
  hgext/rebase.py

CHANGE DETAILS




To: phillco, #hg-reviewers
Cc: mercurial-devel
phabricator - Dec. 8, 2017, 6:45 p.m.
durin42 added a comment.


  @dlax I see a request changes here, but I don't see any commentary as to why?

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D1616

To: phillco, #hg-reviewers, dlax
Cc: durin42, mercurial-devel
phabricator - Dec. 8, 2017, 7:33 p.m.
dlax added a comment.


  In https://phab.mercurial-scm.org/D1616#27819, @durin42 wrote:
  
  > @dlax I see a request changes here, but I don't see any commentary as to why?
  
  
  I meant to comment that the first hunks (i.e.  those above the big comment block) seem unrelated and to ask for clarification.
  No idea where my comment went, sorry...

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D1616

To: phillco, #hg-reviewers, dlax
Cc: durin42, mercurial-devel
phabricator - Dec. 8, 2017, 8:07 p.m.
phillco added a comment.


  @dlax you mean the parts where we pass `opts` instead of `inmemory`? This is necessary for the function to mutate `opts['inmemory']`.

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D1616

To: phillco, #hg-reviewers, dlax
Cc: durin42, mercurial-devel

Patch

diff --git a/hgext/rebase.py b/hgext/rebase.py
--- a/hgext/rebase.py
+++ b/hgext/rebase.py
@@ -179,7 +179,6 @@ 
         self.keepopen = opts.get('keepopen', False)
         self.obsoletenotrebased = {}
         self.obsoletewithoutsuccessorindestination = set()
-        self.inmemory = opts.get('inmemory', False)
 
     @property
     def repo(self):
@@ -830,7 +829,8 @@ 
         else:
             destmap = _definedestmap(ui, repo, destf, srcf, basef, revf,
                                      destspace=destspace,
-                                     inmemory=opts['inmemory'])
+                                     opts=opts)
+            rbsrt.inmemory = opts['inmemory']
             retcode = rbsrt._preparenewrebase(destmap)
             if retcode is not None:
                 return retcode
@@ -850,7 +850,7 @@ 
         rbsrt._finishrebase()
 
 def _definedestmap(ui, repo, destf=None, srcf=None, basef=None, revf=None,
-                   destspace=None, inmemory=False):
+                   destspace=None, opts=None):
     """use revisions argument to define destmap {srcrev: destrev}"""
     if revf is None:
         revf = []
@@ -864,7 +864,7 @@ 
     if revf and srcf:
         raise error.Abort(_('cannot specify both a revision and a source'))
 
-    if not inmemory:
+    if not opts['inmemory']:
         cmdutil.checkunfinished(repo)
         cmdutil.bailifchanged(repo)
 
@@ -939,6 +939,22 @@ 
                 ui.status(_('nothing to rebase from %s to %s\n') %
                           ('+'.join(str(repo[r]) for r in base), dest))
             return None
+    # If rebasing the working copy parent, force in-memory merge to be off.
+    #
+    # This is because the extra work of checking out the newly rebased commit
+    # outweights the benefits of rebasing in-memory, and executing an extra
+    # update command adds a bit of overhead, so better to just do it on disk. In
+    # all other cases leave it on.
+    #
+    # Note that there are cases where this isn't true -- e.g., rebasing large
+    # stacks that include the WCP. However, I'm not yet sure where the cutoff
+    # is.
+    rebasingwcp = repo['.'].rev() in rebaseset
+    if opts['inmemory'] and rebasingwcp:
+        opts['inmemory'] = False
+        # Check these since we did not before.
+        cmdutil.checkunfinished(repo)
+        cmdutil.bailifchanged(repo)
 
     if not destf:
         dest = repo[_destrebase(repo, rebaseset, destspace=destspace)]