[release/3.1] Port VSP floating-point hang from 4.8 #5020
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ghost
requested review from
SamBent
changed the title
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Addresses Issue #4978. This is a port of a servicing fix in .NET 4.7-4.8.
Description
The hang is due to floating-point drift. After any scroll, VSP runs logic to verify that it scrolled to the right place (it can be wrong due to new information discovered while de-virtualizing). There's a special case when the scroll reaches the end of the data, comparing the actual offset to the maximum (extent - viewport height). This uses the strict test DoubleUtil.AreClose(actual, maximum), which tolerates floating-point drift of 10^-16 pixel. But the difference in the repro is 10^-14 pixel, so the test fails.
Prior to the 2020-08 update, this would have merely triggered a Debug.Assert, invisible to customers. With the update, it now requests a re-measure, fixing a bug that arose when the numbers were legitimately different. But when the numbers are so close and layout rounding is enabled, the re-measure just ends up in the same situation, and the app hangs.
The fix is to use the looser test LayoutDoubleUtil.AreClose(actual, maximum), which tolerates floating-point drift of 10^-6 pixel.
Customer Impact
Hang while scrolling.
Regression
Exposed by PR #3564.
Testing
Ad-hoc with customer scenarios.
Standard regression testing.
Risk
Low. Straightforward port of .NETFx fix that was released earlier this year.