Bug 1254815 - [abrt] cmake: operator new(): cmake killed by SIGSEGV
[abrt] cmake: operator new(): cmake killed by SIGSEGV
Status: CLOSED WONTFIX
Product: Fedora
Classification: Fedora
Component: cmake (Show other bugs)
21
x86_64 Unspecified
unspecified Severity unspecified
: ---
: ---
Assigned To: Orion Poplawski
Fedora Extras Quality Assurance
https://retrace.fedoraproject.org/faf...
abrt_hash:0cce62ac62af0413e65f1edadf1...
:
Depends On:
Blocks:
  Show dependency treegraph
 
Reported: 2015-08-18 19:08 EDT by dashesy
Modified: 2015-08-19 11:22 EDT (History)
6 users (show)

See Also:
Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Environment:
Last Closed: 2015-08-19 11:22:50 EDT
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---


Attachments (Terms of Use)
File: cgroup (190 bytes, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: core_backtrace (87.50 KB, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: dso_list (3.63 KB, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: environ (5.46 KB, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: exploitable (82 bytes, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: limits (1.29 KB, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: maps (18.46 KB, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: open_fds (140 bytes, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: proc_pid_status (967 bytes, text/plain)
2015-08-18 19:08 EDT, dashesy
no flags Details
File: backtrace (16.00 MB, text/plain)
2015-08-18 19:09 EDT, dashesy
no flags Details

  None (edit)
Description dashesy 2015-08-18 19:08:36 EDT
Description of problem:
ran cmake for dlib project after addign a FindOpenCV.cmake

Version-Release number of selected component:
cmake-3.0.2-2.fc21

Additional info:
reporter:       libreport-2.3.0
backtrace_rating: 4
cmdline:        cmake ..
crash_function: operator new
executable:     /usr/bin/cmake
kernel:         4.0.7-200.fc21.x86_64
runlevel:       N 5
type:           CCpp
uid:            1000

Potential duplicate: bug 1035624
Comment 1 dashesy 2015-08-18 19:08:38 EDT
Created attachment 1064563 [details]
File: cgroup
Comment 2 dashesy 2015-08-18 19:08:38 EDT
Created attachment 1064564 [details]
File: core_backtrace
Comment 3 dashesy 2015-08-18 19:08:39 EDT
Created attachment 1064565 [details]
File: dso_list
Comment 4 dashesy 2015-08-18 19:08:40 EDT
Created attachment 1064566 [details]
File: environ
Comment 5 dashesy 2015-08-18 19:08:40 EDT
Created attachment 1064567 [details]
File: exploitable
Comment 6 dashesy 2015-08-18 19:08:41 EDT
Created attachment 1064568 [details]
File: limits
Comment 7 dashesy 2015-08-18 19:08:42 EDT
Created attachment 1064569 [details]
File: maps
Comment 8 dashesy 2015-08-18 19:08:42 EDT
Created attachment 1064570 [details]
File: open_fds
Comment 9 dashesy 2015-08-18 19:08:43 EDT
Created attachment 1064571 [details]
File: proc_pid_status
Comment 10 dashesy 2015-08-18 19:09:01 EDT
Created attachment 1064572 [details]
File: backtrace
Comment 11 Orion Poplawski 2015-08-18 21:17:56 EDT
Can you give a bit more details of what you did?  Looks like there may be some king of recursion going on, perhaps by having the FindOpenCV.cmake file in the out of tree build directory?
Comment 12 dashesy 2015-08-18 21:26:45 EDT
Yes it could be. OpenCV cmake file was not found, so I copied a version of it to the `build` directory and then exported the `build` directory to `CMAKE_MODULE_PATH` (and in my tries to see what works I think also `CMAKE_PREFIX_PATH`).
Comment 13 Orion Poplawski 2015-08-18 21:29:20 EDT
Yeah, you really want to start with an empty build directory.  No source files there.  Let me know how that goes.
Comment 14 dashesy 2015-08-18 21:33:08 EDT
It works that way :) but is a core dump normal otherwise? i.e. is there no other way to avoid the recursion, I had never seen cmake crash.
Comment 15 Orion Poplawski 2015-08-19 11:22:50 EDT
Obviously a core dump is not ideal, but your usage was pretty far outside of normal as well.  So I'm not sure it's worth spending too much more time on it.

Note You need to log in before you can comment on or make changes to this bug.