Open xiaxinmeng opened 1 year ago
This appears to be memory exhaustion.
Not sure about the funny span errors - that is a go internal error.
fatal error: found bad pointer in Go heap (incorrect use of unsafe or cgo?)
That looks like a a bug to report upstream to the go project. It would need a reproducer though that doesn't involve gpython preferably!
On Wed Dec 7, 2022 at 16:54 CET, Nick Craig-Wood wrote:
This appears to be memory exhaustion.
Not sure about the funny span errors - that is a go internal error.
fatal error: found bad pointer in Go heap (incorrect use of unsafe or cgo?)
That looks like a a bug to report upstream to the go project. It would need a reproducer though that doesn't involve gpython preferably!
isn't this either:
(running this locally w/ Go-1.19, I "just" killed my machine by OOM)
-s
It's basically an OOM problem I think, but the wasm code behaved very strangely.
It would be nice to verify this with a fresh gpython compile on wasm latest.
Taking long operations as argument of compiler() leads that pointer gets access to unused region crashing gpython
test.py
compile('1'+'<2'*100000000,'','exec')
Output on go/wasm(https://gpython.org/?wasm):