[pypy-issue] Issue #2035: Loop header in PYPYLOG may miss get_printable_location result in some cases (pypy/pypy)

Tobias Pape issues-reply at bitbucket.org
Mon Apr 27 20:58:31 CEST 2015


New issue 2035: Loop header in PYPYLOG may miss get_printable_location result in some cases
https://bitbucket.org/pypy/pypy/issue/2035/loop-header-in-pypylog-may-miss

Tobias Pape:

In [RSqueak](https://github.com/HPI-SWA-Lab/RSqueak), sometimes traces have the right `get_printable_location` result in a `debug_merge_point` but not in the Loop label:

(see [Gist](https://gist.github.com/krono/18131648ba6dd7d7184c))
```
…
# Loop 0 ((jitdriver: get_printable_location disabled, no debug_print)) : entry bridge with 65 ops
…
debug_merge_point(0, 0, '(ByteString class >> #stringHash:initialHash:) [9]: <0x15>pushTemporaryVariableBytecode(5)')
…
debug_merge_point(0, 0, '(ByteString class >> #stringHash:initialHash:) [12]: <0xac>longJumpIfFalseBytecode(0)')
debug_merge_point(0, 0, '(ByteString class >> #stringHash:initialHash:) [51]: <0x13>pushTemporaryVariableBytecode(3)')
debug_merge_point(0, 0, '(ByteString class >> #stringHash:initialHash:) [52]: <0x7c>returnTopFromMethodBytecode')
…
```




More information about the pypy-issue mailing list