You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have been experiencing trouble with the most recent versions of psalm.
We're currently running on 5.7.7, but we walked back recent upgrades, and it appears that 5.2 does not have the problem, but our upgrade to 5.6 seems to bring back the problem. This may also be related to the addition of UnusedBaselineEntry, which we've recently enabled. Version 5.5 seems to be the relevant release that changed this behavior for us.
Specifically, if we have duplicated <code> blocks in our baseline, fixing only one of them does not correctly remove the second instance from the baseline as expected.
We have been experiencing trouble with the most recent versions of psalm.
We're currently running on 5.7.7, but we walked back recent upgrades, and it appears that 5.2 does not have the problem, but our upgrade to 5.6 seems to bring back the problem. This may also be related to the addition of
UnusedBaselineEntry
, which we've recently enabled. Version 5.5 seems to be the relevant release that changed this behavior for us.Also, removing the
occurrence
attribute also may have exposed the problem.It appears that the ErrorBaseline code:
psalm/src/Psalm/ErrorBaseline.php
Lines 140 to 147 in 9cf4f60
as well as the test cases, are concerned still with absolute number of occurrences. This could be confusing the issue.
We have a test case created that seems to reproduce the problem we are seeing:
iFixit@477e0a2
Notice that in our example/testcase we are avoiding encoding the
occurrence
s field in the example baseline XML:https://github.com/iFixit/psalm/blob/477e0a26bb8dff7772f055da83343909af686679/tests/ErrorBaselineTest.php#L498
but that the remain test cases are still using
occurrences
.Specifically, if we have duplicated
<code>
blocks in our baseline, fixing only one of them does not correctly remove the second instance from the baseline as expected.Please advise.
CC @jarstelfox @ardelato @sctice-ifixit
The text was updated successfully, but these errors were encountered: