D7905: rebase: move some variables after an error cases where they're not needed

martinvonz (Martin von Zweigbergk) phabricator at mercurial-scm.org
Fri Jan 24 18:27:29 EST 2020


Closed by commit rHG1cb7ae9b0071: rebase: move some variables after an error cases where they're not needed (authored by martinvonz).
This revision was automatically updated to reflect the committed changes.

REPOSITORY
  rHG Mercurial

CHANGES SINCE LAST UPDATE
  https://phab.mercurial-scm.org/D7905?vs=19341&id=19579

CHANGES SINCE LAST ACTION
  https://phab.mercurial-scm.org/D7905/new/

REVISION DETAIL
  https://phab.mercurial-scm.org/D7905

AFFECTED FILES
  hgext/rebase.py

CHANGE DETAILS

diff --git a/hgext/rebase.py b/hgext/rebase.py
--- a/hgext/rebase.py
+++ b/hgext/rebase.py
@@ -1793,14 +1793,6 @@
             for i, revs in enumerate(unwanted)
             if revs is not None
         )
-        base = bases[i]
-
-        # newps[0] should match merge base if possible. Currently, if newps[i]
-        # is nullrev, the only case is newps[i] and newps[j] (j < i), one is
-        # the other's ancestor. In that case, it's fine to not swap newps here.
-        # (see CASE-1 and CASE-2 above)
-        if i != 0 and newps[i] != nullrev:
-            newps[0], newps[i] = newps[i], newps[0]
 
         # The merge will include unwanted revisions. Abort now. Revisit this if
         # we have a more advanced merge algorithm that handles multiple bases.
@@ -1817,6 +1809,15 @@
                 % (rev, repo[rev], unwanteddesc)
             )
 
+        base = bases[i]
+
+        # newps[0] should match merge base if possible. Currently, if newps[i]
+        # is nullrev, the only case is newps[i] and newps[j] (j < i), one is
+        # the other's ancestor. In that case, it's fine to not swap newps here.
+        # (see CASE-1 and CASE-2 above)
+        if i != 0 and newps[i] != nullrev:
+            newps[0], newps[i] = newps[i], newps[0]
+
     repo.ui.debug(b" future parents are %d and %d\n" % tuple(newps))
 
     return newps[0], newps[1], base



To: martinvonz, #hg-reviewers, pulkit
Cc: mercurial-devel


More information about the Mercurial-devel mailing list