golden/test fails on Scala 2.13.6
#103
Comments
|
Thanks for the report. I run into this from time to time and I think this
is due to differences in Java2D rendering between JVM versions and OS
versions. What JVM version does the 2.13.6 build use, and, if Doodle is in
the 3 build, does it differ from that? Thanks!
…On Sun, Aug 29, 2021 at 4:26 PM Seth Tisue ***@***.***> wrote:
++2.13.6 followed by golden/test gives
==> X doodle.golden.Text.text-color 0.045s munit.ComparisonFailException: /Users/tisue/doodle/golden/src/test/scala/doodle/golden/Text.scala:40
39:
40: testPicture("text-color") {
41: text[Algebra, Drawing]("Red")
Heights differ
=> Obtained
34
=> Diff (- obtained, + expected)
-34
+33
at munit.FunSuite.assertEquals(FunSuite.scala:11)
at doodle.golden.GoldenPicture.assertGoldenPicture(GoldenPicture.scala:32)
at doodle.golden.GoldenPicture.assertGoldenPicture$(GoldenPicture.scala:16)
at doodle.golden.Text.assertGoldenPicture(Text.scala:11)
at doodle.golden.GoldenPicture.$anonfun$testPicture$1(GoldenPicture.scala:59)
this came up in the Scala 2.13 community build
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#103>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAAH5XFNDEGKBNJ2A7W5CCDT7JGRDANCNFSM5DALA4AA>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
|
|
I am not sure if I have seen it fail on our Jenkins (which runs Ubuntu) or not. But I know I'm able to reproduce it locally on macOS 11.5.2 on GraalVM 11: and also AdoptOpenJDK 8: |
|
So far I've been treating the CI results (Github actions on Ubuntu) for the golden tests as definitive, and mostly ignoring what happens on my local machine beyond visual inspection of the generated images. I could make the image diffing more robust to differences in size. In general image diffing is a hard problem, and I'm not sure how much effort to invest in this. |
|
I've added some code that might address this. |
|
it still fails for me. (after #104, which is needed for the tests to even compile.) |
|
It does fail with a different error message, though, right? I can make the code print out the offending image and the diff, which might go some way to debugging this. I fear writing an acceptable image diff algorithm may become a project in its own right. |
|
currently the error is: I think you could reasonably mark the test as to be skipped except on Linux. |
|
I've updated the error message so there should hopefully be much more useful output. If you can paste that in we'll be able to tell if its a spurious error or not, and take action accordingly. |
|
AdoptOpenJDK Java 1.8.0_292, macOS: |
|
Thanks. It looks like a small difference in font rendering. I've attempted to make the test more robust to this difference. If this doesn't work I'll mark the test as Linux only. |
|
still failing |
|
Ok. I've made my final attempt. Let me know if it works... |
🥇 |
|
😹 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
++2.13.6followed bygolden/testgivesthis came up in the Scala 2.13 community build
The text was updated successfully, but these errors were encountered: