Skip to content

Conversation

@slavek-kucera
Copy link
Contributor

While migrating to 4.0.19 I discovered that enabling optimizations removes ParenthesizedExpression node from AST causing the littleEndianHeap to not perform the transform correctly.

@slavek-kucera slavek-kucera marked this pull request as ready for review November 14, 2025 11:09
@sbc100 sbc100 requested a review from RReverser November 14, 2025 16:56
@sbc100
Copy link
Collaborator

sbc100 commented Nov 14, 2025

Sorry if I've asked you this already, but what is you use case for bigendian suport @slavek-kucera ? Are you running in production on actual BE machines?

@slavek-kucera
Copy link
Contributor Author

Yes, the output targets s390x running z/os.

@slavek-kucera
Copy link
Contributor Author

As for testing, I guess we could add bigendian3 test suite (assuming @juj still runs it in the private pipeline).

@juj
Copy link
Collaborator

juj commented Nov 18, 2025

As for testing, I guess we could add bigendian3 test suite (assuming @juj still runs it in the private pipeline).

I totally do. (e.g. http://clbri.com:8010/#/builders/11/builds/1275 - ignore lots of red for other failures, but there's a passing bigendian0 at the very bottom of the page)

Posted PR #25820 to complete the testing matrix. After that lands, I can add the remaining bigendian suites onto the test runner list.

@sbc100
Copy link
Collaborator

sbc100 commented Nov 18, 2025

Yes, the output targets s390x running z/os.

What engine are you running the emscripten output on? i.e. Are these servers running Nodejs I guess?

@slavek-kucera
Copy link
Contributor Author

Yes, the engine is nodejs.

juj added a commit that referenced this pull request Nov 20, 2025
Add the rest of the bigendian test matrix to help test
#25797
@juj
Copy link
Collaborator

juj commented Nov 20, 2025

@slavek-kucera My Linux box now completed run of all the bigendian suites. It looks like there are some failures in -O2 and -Os modes: http://clbri.com:8010/#/builders/11/builds/1278

@slavek-kucera
Copy link
Contributor Author

I think those are all false positive build failures due to closure compiler.

Function LE_ATOMICS_WAITASYNC: called with 3 argument(s). Function requires at least 4 argument(s) and no more than 4 argument(s)

Maybe the function needs an annotation just like Atomics.waitAsync?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants