View Issue Details Jump to Notes ] Print ]
IDProjectCategoryView StatusDate SubmittedLast Update
0013258CMakeCMakepublic2012-05-30 11:292012-12-03 07:46
ReporterClinton Stimpson 
Assigned ToDavid Cole 
StatusclosedResolutionno change required 
PlatformOSOS Version
Product VersionCMake 2.8.8 
Target VersionCMake 2.8.9Fixed in VersionCMake 2.8.9 
Summary0013258: valgrind reports memory problems in cmIfCommand.cxx
DescriptionUsing DeployQt4.cmake, I can get some odd behavior.
Steps To Reproducefind_package(Qt4)
add_executable(myapp myapp.cpp)
target_link_libraries(myapp ${QT_LIBRARIES})
install(TARGETS myapp DESTINATION .)
install_qt4_executable(myapp "qtiff;qico")

Run cmake. No build or install necessary to see the problem.
Additional InformationValgrind reports:

==30955== Invalid read of size 8
==30955== at 0x55DA147: __GI___strncasecmp_l (strcmp.S:215)
==30955== by 0x5590EDA: ____strtod_l_internal (strtod_l.c:566)
==30955== by 0x711377: (anonymous namespace)::GetBooleanValue(std::string&, cmMakefile*) (cmIfCommand.cxx:242)

That's triggered by the if(${executable_path}) at DeployQt4.cmake line 101, and it appears that executable_path is populated with an unrelated variable.
TagsNo tags attached.
Attached Files


David Cole (manager)
2012-06-13 16:53

I changed the dashboard script that drives the CMake valgrind dashboard submission. I added building the Qt cmake-gui, which should have the side effect of activating the Qt4Deploy test on that dashboard.

We'll see if that causes this valgrind report, too.

If not, I'll add the code reported here as-is to the test suite to try to reproduce this...

Then we can figure out the right fix from there.
Mike McQuaid (reporter)
2012-07-03 19:14

Any update on the dashboard valgrind report? Thanks!
Clinton Stimpson (developer)
2012-07-03 19:23

I was recently debugging a BundleUtitities test failure on another Linux machine, and saw these valgrind errors again. A google search of "valgrind strtod" hints that its a false positive.

I can't remember the original problem off the top of my head, but maybe you want to look at this bug report and see if the -pie linker flag is the real cause of your problems that led to this bug report. [^]
David Cole (manager)
2012-07-09 07:13

Unset target version field; too late for a fix to go into 2.8.9; deferred until a future version.
Mike McQuaid (reporter)
2012-07-09 07:18

In that case can my DeployQt4 workaround patch go in? Clinton, thoughts?
Clinton Stimpson (developer)
2012-07-09 12:20

I think we can close this bug, and mark it invalid. I think the valgrind error is a false positive, and it appears unrelated to Mike's original problem.

Here's a better report of the original. [^]
David Cole (manager)
2012-07-09 12:52

Resolving as "no change required" per the discussion in the recent notes.
David Cole (manager)
2012-12-03 07:46

Closing resolved issues that have not been updated in more than 4 months.

 Issue History
Date Modified Username Field Change
2012-05-30 11:29 Clinton Stimpson New Issue
2012-06-05 14:46 David Cole Assigned To => David Cole
2012-06-05 14:46 David Cole Status new => assigned
2012-06-13 16:53 David Cole Note Added: 0029689
2012-06-13 16:54 David Cole Target Version => CMake 2.8.9
2012-07-03 19:14 Mike McQuaid Note Added: 0029901
2012-07-03 19:23 Clinton Stimpson Note Added: 0029902
2012-07-09 07:13 David Cole Note Added: 0029986
2012-07-09 07:18 Mike McQuaid Note Added: 0029993
2012-07-09 07:19 David Cole Target Version CMake 2.8.9 =>
2012-07-09 12:20 Clinton Stimpson Note Added: 0030005
2012-07-09 12:52 David Cole Note Added: 0030008
2012-07-09 12:52 David Cole Status assigned => resolved
2012-07-09 12:52 David Cole Resolution open => no change required
2012-07-09 12:52 David Cole Product Version => CMake 2.8.8
2012-07-09 12:52 David Cole Fixed in Version => CMake 2.8.9
2012-07-09 12:52 David Cole Target Version => CMake 2.8.9
2012-12-03 07:46 David Cole Note Added: 0031809
2012-12-03 07:46 David Cole Status resolved => closed

Copyright © 2000 - 2018 MantisBT Team