New issue
Advanced search Search tips

Issue 753851 link

Starred by 2 users

Issue metadata

Status: Fixed
Owner:
Closed: Jun 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Fuchsia
Pri: 3
Type: Bug

Blocking:
issue 706592



Sign in to add a comment

Write a test for LoggingTest.CheckCausesDistinctBreakpoints on Fuchsia

Project Member Reported by scottmg@chromium.org, Aug 9 2017

Issue description

I disabled LoggingTest.CheckCausesDistinctBreakpoints here https://chromium-review.googlesource.com/c/607410 on Fuchsia.

Rather than being lazy about it, it could be reimplemented by doing something other than fork() to communicate info back and forth between parent and child.
 

Comment 1 by w...@chromium.org, Aug 9 2017

Owner: w...@chromium.org
Status: Started (was: Available)

Comment 2 by w...@chromium.org, Aug 12 2017

We can switch the child process to be launched via the multi-process-test
macro, but the bulk of the work will be actually capturing the CHECK
crashes; once we have that we can also enable GTEST death-tests if we want.

Looks like we can crib from Fuchsia's ulib unit-test crash handler
implementation (see
https://github.com/fuchsia-mirror/magenta/blob/82b9b705e91b798f35ed1daf831ca2bcad6689bd/system/ulib/unittest/crash-handler.c#L317
).

Comment 3 by w...@chromium.org, Aug 15 2017

Status: Assigned (was: Started)
Punting this back to Assigned, since it's going to take more work than just updating process-launching. We'll want to set up an "exception port" for the crashy child process to trap it exploding, as per: https://fuchsia.googlesource.com/magenta/+/master/docs/syscalls/task_bind_exception_port.md

Comment 4 by w...@chromium.org, Apr 7 2018

Owner: ----
Status: Available (was: Assigned)
Assigning to fdegans@, who is adding the 
Owner: fdegans@chromium.org
Status: Assigned (was: Available)
Actually assigning myself. wez@ got captured by Candlejack before he could finish his me
Cc: scottmg@chromium.org
I'm looking into this test and I'm a bit confused by what it does. It seems it is checking that the platform exception handling is working properly. I'm not sure what part of Chromium is under test here.

Comment 7 by w...@chromium.org, May 24 2018

Re #6: The test is intended to verify that CHECK()s at different call-sites can be distinguished, rather than all being folded into a single set of instructions that the CHECK sites jump to on failure, IIUC.
Project Member

Comment 8 by bugdroid1@chromium.org, Jun 8 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a2a5386474feb62d5b8a84244b8657f09209fb22

commit a2a5386474feb62d5b8a84244b8657f09209fb22
Author: Fabrice de Gans-Riberi <fdegans@chromium.org>
Date: Fri Jun 08 18:04:26 2018

Implement LoggingTest.CheckCausesDistinctBreakpoints for Fuchsia.

Bug:  753851 
Change-Id: Id457ec00228fe6a5592719c47491087e979bd5b2
Reviewed-on: https://chromium-review.googlesource.com/1077600
Commit-Queue: Fabrice de Gans-Riberi <fdegans@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Reviewed-by: Scott Graham <scottmg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#565682}
[modify] https://crrev.com/a2a5386474feb62d5b8a84244b8657f09209fb22/base/logging_unittest.cc

Status: Fixed (was: Assigned)

Sign in to add a comment