See the explanation for more.
Details
- Reviewers
martinvonz - Group Reviewers
hg-reviewers - Commits
- rHG34a8ef358c93: test-dirstate-race: hide irrelevant hg status output
Diff Detail
- Repository
- rHG Mercurial
- Lint
Lint Skipped - Unit
Unit Tests Skipped
Event Timeline
tests/test-dirstate-race.t | ||
---|---|---|
223–231 | The commit that introduced this test case (15e85dded933) referred to bug https://bz.mercurial-scm.org/show_bug.cgi?id=5581. The description of that issue (filed by you) finishes with "That last 'hg status' should be empty, but it returns 'M b'.". You changed your mind about what the expected behavior should be? |
tests/test-dirstate-race.t | ||
---|---|---|
223–231 | The *last* hg status should be empty. This is an intermediate hg status. |
tests/test-dirstate-race.t | ||
---|---|---|
244 | ^^ this one should be empty. |
tests/test-dirstate-race.t | ||
---|---|---|
223–231 | Oh, the "hg status" is just to cause it to write the dirstate? Would "hg debugrebuilddirstate" work? That doesn't produce any output and it's clearer that that will result in the dirstate getting written ("hg status" doesn't always write it, as I'm sure you know). If that would work, we can delete the whole comment, I think. |
tests/test-dirstate-race.t | ||
---|---|---|
223–231 | I'm afraid not -- debugrebuilddirstate will try to acquire the wlock, which is already held by the rebase. |
tests/test-dirstate-race.t | ||
---|---|---|
223–231 | Makes sense (for others' info: "hg status" will write the dirstate if the lock can be acquired, but will not wait for the lock, IIRC). How about piping the result of that "hg status" to /dev/null and adding a comment above it saying that we're only calling it to cause the dirstate to be written if the repo is not locked? Then you can get rid of the comment above. |
The commit that introduced this test case (15e85dded933) referred to bug https://bz.mercurial-scm.org/show_bug.cgi?id=5581. The description of that issue (filed by you) finishes with "That last 'hg status' should be empty, but it returns 'M b'.". You changed your mind about what the expected behavior should be?