rdamazio added a comment.
The flakiness was just due to Python map order randomization in tests (a fixed PYTHONHASHSEED=3170796678 or similar reproduces it consistently by getting dln before lognull). The real issue was the added "ignoresub=True" argument which broke strict ordering guarantees in the return value, combined with the fact that each lazy alias entry makes a copy of the cmdtable *at the state when it's created*. I've fixed this in ui.configitems so it preserves ordering - PTAL. Should also be a tiny bit faster now. I've also added tests for the recursive alias case. REPOSITORY rHG Mercurial REVISION DETAIL https://phab.mercurial-scm.org/D2678 To: rdamazio, #hg-reviewers, durin42, pulkit Cc: mharbison72, spectral, pulkit, mercurial-devel _______________________________________________ Mercurial-devel mailing list Mercurial-devel@mercurial-scm.org https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
