Updates:
Owner: [email protected]
Comment #8 on issue 2921 by [email protected]:
mjsunit/debug-step-4-in-frame is flaky and slow
http://code.google.com/p/v8/issues/detail?id=2921
I have reduced this to a problem with the interaction between the
optimizing compiler and the debugger. If the test is run with
hydrogen_filter=g, the test fails (throws an exception), if it is run with
a filter excluding "g", it succeeds. Handing off to Yang to track down
further, since this is indeed V8 debugger black voodoo magic at work.
The attached .js file trims out unrelated aspects of the test and just
illustrates the problem if run like this:
out/x64.debug/d8 --test --stress-opt --always-opt --nobreak-on-abort
--nodead-code-elimination --nofold-constants --expose-debug-as debug
test/mjsunit/mjsunit.js test/mjsunit/debug-step-4-in-frame-edit.js
--hydrogen-filter=g
The problem repros on all platforms.
It seems that the first "Here" message for stress run 2 is already wrong
and shows the first sign of trouble even before the exception is thrown:
============ Stress 2/2 ============
Test case, parameters undefined/0
Here (1/0): for (state[2] = 0; state[2] < 5; state[2]++) {
In successful runs it reports a different source line:
============ Stress 2/2 ============
Test case, parameters undefined/0
Here (1/0): var a = 1978;
Attachments:
debug-step-4-in-frame-edit.js 3.6 KB
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
