Remove reliance on "questionable" UpdateFile operation#52441
Remove reliance on "questionable" UpdateFile operation#52441sharwell merged 1 commit intodotnet:mainfrom
Conversation
|
What is making UpdateFile unreliable? Should we remove it? |
No idea. It's just a guess based on logs like this:
I wanted to remove it, but it's still used by another test. |
jasonmalinowski
left a comment
There was a problem hiding this comment.
I guess if we think this is unsafe to use, then we should remove it and update tests. Or we don't have the evidence this is actually problematic then we should clarify the PR title accordingly.
|
@jasonmalinowski The code is used in three locations. This is the only location where tests in the last 30 days failed to achieve the desired outcome following the call. The issue has not been locally reproducible, and we have no evidence suggesting the reason why this call fails at this location. This pull request converts an operation which does not always produce the expected results (unreliable) to another form used elsewhere that does appear to produce the expected results. |
|
@jasonmalinowski I believe this addresses some of the issues reported in #52409 |

During investigation of #52409, logs were observed where the editor text did not appear to match the text set up by
InitializeAsync(e.g. #52441 (comment) below). Since theInitializeAsyncmethod set the text using a somewhat unique method, and other tests aren't showing signs of incorrect values, we modify the problematic test to set the editor text using the same coding pattern that other non-problematic tests are using.The failure has not been locally reproducible, so it's not clear why (or to some degree if) the
UpdateFilecall is failing in this scenario.The new code follows the approach of
AbstractEditorTest.SetUpEditor.Closes #52386