D20547 alone would also fix this specific problem (by skipping the Herald transactions), but I think this change is good on its own and generally makes behavior more consistent/expected.

Reading PHI1222; nice catch!

Currently, I think the primary transactions and transactions triggered by Herald are getting different group IDs. They should be the same group ID, e.g. all these transactions should be in the same "group" for the purposes of collecting transactions into effect groups:

(This is mostly for my own notes and contains roughly zero insight.)

If we were just doing personal notification sounds, I think there'd be a good argument for making this a Settings panel.

Thanks, T13293 makes this much clearer.

Per D20533, the major query this UI uses is currently unkeyed (no dateCreated key on transaction tables).

See also T13290. Now that full URIs count as mentions (D20527), the text "Differential Revision: ..." in commits is counting as a mention, and "mentioning" the associated revision.

(We also don't need to tick the y-max -- we could tick at 100 and draw to 110.)

Yeah, I ideally want to try to round up, up to X%, to the nearest "nice" number, like if the range is 0-93 we show 0-100, since "100" is the nearest "nice round number" above 93. There was some simple code for this before (visual-max = actual-max * 1.2 or something) which makes the shape of the graph look good but doesn't produce the most satisfying ticks on the Y-axis or the ideally satisfying Y-max.

Implement parseVersionOutput in ArcanistExternalLinter so that it can be proxied from within another instance of ArcanistExternalLinter

Rebase

Rebase

Unrelated but I noticed that ArcanistBundleTestCase::testGitRepository fails locally for me when running arc unit --everything: