Uh oh!
There was an error while loading. Please reload this page.
report: do not use uv_default_loop() as fallback - #25652
Conversation
addaleax
commented
Jan 23, 2019
There was a problem hiding this comment.
Since the report is meant to aid diagnostics, is there any value in logging that env was null in the report?
There was a problem hiding this comment.
@richardlau I guess we could do that – do you have any suggestion on how to represent that?
gireeshpunathil
commented
Jan 23, 2019
IIRC, under
Something to investigate in future, but irrespective of that, this change looks good to me, for a clean and meaningful report. |
addaleax
commented
Jan 23, 2019
@gireeshpunathil I think the test that checks I think for other fatal errors (e.g. calling |
gireeshpunathil
commented
Jan 24, 2019
@addaleax - thanks for the clarification, I wasn't aware of such |
Not seeing an associated `Environment` is a rare condition anyway, but using `uv_default_loop()` as a fallback is not thread-safe.
0bfd3bc to
0542913Compareuv_default_loop()uv_default_loop() as fallbackaddaleax
commented
Jan 24, 2019
(If you’re wondering why I changed the commit message, it just occurred to me that the previous one is kind of ableist.) |
addaleax
commented
Jan 25, 2019
Trott
commented
Jan 26, 2019
addaleax
commented
Jan 27, 2019
Landed in bb774b1 |
Not seeing an associated `Environment` is a rare condition anyway, but using `uv_default_loop()` as a fallback is not thread-safe. PR-URL: #25652 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Not seeing an associated `Environment` is a rare condition anyway, but using `uv_default_loop()` as a fallback is not thread-safe. PR-URL: #25652 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Not seeing an associated
Environmentis a rare condition anyway,but using
uv_default_loop()as a fallback is not thread-safe.Checklist
make -j4 test(UNIX), orvcbuild test(Windows) passes