[Merged by Bors] - Correct pop_on_return behaviour #1853
Conversation
Codecov Report
@@ Coverage Diff @@
## main #1853 +/- ##
=======================================
Coverage 56.29% 56.30%
=======================================
Files 201 201
Lines 17936 17937 +1
=======================================
+ Hits 10097 10099 +2
+ Misses 7839 7838 -1
Continue to review full report at Codecov.
|
Razican
requested review from
Razican,
jasonwilliams,
raskad,
HalidOdat,
RageKnify and
jedel1043
There was a problem hiding this comment.
Feels weird to not check anything after executing the JS code. We could at least check whether an exception was thrown, in the first case the result of exec should be Err and in the second case Ok, right? Just asserting that they got the right variants feels better than doing nothing else.
|
Sorry for the late response on this; busy Monday. :) Do the changes provided suffice? |
This PR changes the following: - Updates the value of `pop_on_return` after a catch as to prevent VM stack corruption - Adds two test cases which demonstrate the issue and demonstrate that it has been fixed I am unsure if it is possible to abuse the patch provided; one would need to catch from within an array initialisation without calling into another frame (e.g., with a lambda), which I don't think is possible.
|
Pull request successfully merged into main. Build succeeded: |
bors
bot
changed the title
Razican
added
bug
This PR changes the following:
pop_on_returnafter a catch as to prevent VM stack corruptionI am unsure if it is possible to abuse the patch provided; one would need to catch from within an array initialisation without calling into another frame (e.g., with a lambda), which I don't think is possible.