Our performance test suite worker detected a regression in debugunbundle (http://perf.octobus.net/#regressions?sort=3&dir=desc): basic_commands.UnbundleTimeSuite.time_debugunbundle('netbeans-2018-08-01', 'last-thousand') 2018-08-17 03:51 d0d197ab..8547c859 1.81x 1.320s 2.394s basic_commands.UnbundleTimeSuite.time_debugunbundle('pypy-2018-08-01', 'last-thousand') 2018-08-17 03:51 d0d197ab..8547c859 1.15x 1.236s 1.419s The bisect tool detected that it's likely because of https://www.mercurial-scm.org/repo/hg/rev/8f83a953. We will send a patch to fix it.
Fixed by https://mercurial-scm.org/repo/hg/rev/a911932d5003 Boris Feld <boris.feld@octobus.net> revlog: reuse cached delta for identical base revision (issue5975) Since 8f83a953dddf, we skip over empty deltas when choosing a delta base. Such delta happens when two distinct revisions have the same content. The remote might be sending a delta against such revision within the bundle. In that case, the delta base is no longer considered, but the cached one could still, be used with the equivalent revision. Not reusing the delta from the bundle can have a significant performance impact, so we now make sure with doing so when possible. (please test the fix)
Bug was set to TESTING for 7 days, resolving