Otherwise the clone command will emit a long stacktrace if there is no =
character.
Details
Diff Detail
- Repository
- rHG Mercurial
- Lint
Automatic diff as part of commit; lint not applicable. - Unit
Automatic diff as part of commit; unit tests not applicable.
Event Timeline
(feel free to put it back in review-needed if my question was silly).
tests/test-clonebundles.t | ||
---|---|---|
458–464 | Your patch is about testing error when only one of the items is bad. Then, your tests only test it with a single item. Am I missing something or should we have a tests with multiple items in the list ? |
tests/test-clonebundles.t | ||
---|---|---|
458–464 | Not sure I understand. There are a handful of tests in test-clonebundles.t that have multiple key/value pairs already, so the good case is covered. (I missed this before and thought that I commented there was no coverage, but I was wrong and don't see that comment now.) The setting needs to be in ui.clonebundles=key=value[,key=value] form. So if it catches a single malformed key/value pair, I'm not sure that it buys us anything to have a good,bad test too, unless I'm missing something. |
tests/test-clonebundles.t | ||
---|---|---|
458–464 | a test mixing good and bad would check that we report error on the appropriate blocks. something like: $ hg --config ui.clonebundleprefers=extra=b,bad,COMPRESSION=unknown clone -U http://localhost:$HGPORT bad-input abort: invalid ui.clonebundleprefers item: bad (each comma separated item should be key=value pairs) [255] |
Your patch is about testing error when only one of the items is bad. Then, your tests only test it with a single item. Am I missing something or should we have a tests with multiple items in the list ?