Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(3779)

Issue 575180043: Issue 5587: allow configure --enable-ubsan (Closed)

Can't Edit
Can't Publish+Mail
Start Review
Created:
4 years, 5 months ago by Dan Eble
Modified:
4 years, 5 months ago
Reviewers:
lemzwerg
CC:
lilypond-devel_gnu.org
Visibility:
Public.

Description

https://sourceforge.net/p/testlilyissues/issues/5587/ The new configuration option --enable-ubsan instruments compiled code with the Undefined Behavior Sanitizer (UBSan). GCC UBSan options are documented here: https://gcc.gnu.org/onlinedocs/gcc-9.2.0/gcc/Instrumentation-Options.html I ran the regression tests with UBSan enabled, and it found that a bool value was being read as something other than 0 or 1. Initializing more of the members of Constrained_breaking fixed that.

Patch Set 1 #

Total comments: 1
Unified diffs Side-by-side diffs Delta from patch set Stats (+46 lines, -14 lines) Patch
M aclocal.m4 View 4 chunks +30 lines, -2 lines 1 comment Download
M lily/constrained-breaking.cc View 3 chunks +15 lines, -11 lines 0 comments Download
M lily/include/constrained-breaking.hh View 1 chunk +1 line, -1 line 0 comments Download

Messages

Total messages: 3
Dan Eble
4 years, 5 months ago (2019-10-25 14:00:52 UTC) #1
lemzwerg
LGTM. Your changes are two commits, right? https://codereview.appspot.com/575180043/diff/559180043/aclocal.m4 File aclocal.m4 (right): https://codereview.appspot.com/575180043/diff/559180043/aclocal.m4#newcode274 aclocal.m4:274: # "print ...
4 years, 5 months ago (2019-10-25 17:08:38 UTC) #2
Dan Eble
4 years, 5 months ago (2019-10-25 20:17:37 UTC) #3
On 2019/10/25 17:08:38, lemzwerg wrote:
> LGTM.  Your changes are two commits, right?

Right.

> aclocal.m4:274: # "print a verbose error report and exit the program"
> Why quotation marks in the comment?

Heh. I'm quoting the GCC manual without citing it.
Sign in to reply to this message.

Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld f62528b