Bug #39917 glibc memory corruption running workbench 5.1.3
Submitted: 7 Oct 2008 22:24 Modified: 29 Jan 2009 12:38
Reporter: Dennis Gilmore Email Updates:
Status: Duplicate Impact on me:
None 
Category:MySQL Workbench Severity:S1 (Critical)
Version:5.1.5 OS:Linux (F10 x86_64)
Assigned to: CPU Architecture:Any

[7 Oct 2008 22:24] Dennis Gilmore
Description:
attempting to run mysql-workbench on a fedora 10 x86_64 system i get a memory corruption error

i will attach the full traceback as a file since I am unable to paste it here.

How to repeat:
run mysql-workbench
[7 Oct 2008 22:25] Dennis Gilmore
traceback running mysql-workbench

Attachment: mysql-workbench-traceback (application/octet-stream, text), 9.24 KiB.

[8 Nov 2008 16:20] Valeriy Kravchuk
Thank you for a problem report. Please, try to repeat with a newer version, 5.1.4, and inform about the results.
[8 Nov 2008 17:15] Dennis Gilmore
traceback running 5.1.4

Attachment: mysql-workbench-5.1.4-traceback (application/octet-stream, text), 61.99 KiB.

[8 Nov 2008 17:18] Dennis Gilmore
5.1.4  gets further.  i get the splash screen then the attached traceback.  the process stays running and the splashscreen stays on the screen.  I end up needing to kill mysql-workbench-bin to get the splashscreen to go
[3 Dec 2008 14:27] Susanne Ebrecht
Many thanks for writing a bug report. The system you are using is unsupported at the moment.

But that doesn't mean that we don't have interests to get feedback about 64bit tries.

How did you install Workbench on your 64bit system?
[4 Dec 2008 2:32] Dennis Gilmore
I installed via rpms that I made.  the 32 bit version did the exact same thing also.
[17 Dec 2008 14:41] Susanne Ebrecht
Many thanks for writing a bug report.

That looks like a depending package is missing.

Please check here if you really have installed all the packages:
http://dev.mysql.com/workbench/?page_id=152

Workbench on Linux is a very new product and it is still in development mode. Our development started here with Linux 32bit and not with Linux 64bit. For 64bit I would recommend you download Fedora 9 Source from

http://dev.mysql.com/downloads/workbench/5.1.html

Then use tar to unzip the package:
$ tar xvzf <package name>

change into the source directory:

$ cd mysql-workbench-5.1.5

know start compiling:
$ ./autogen.sh
$ make
$ make install
[17 Dec 2008 16:01] Dennis Gilmore
spec file that i created for mysql-workbench

Attachment: mysql-workbench.spec (text/x-rpm-spec), 2.64 KiB.

[17 Dec 2008 16:28] Dennis Gilmore
I built the rpm using the spec file that i attached to this bug. it was built using mock  which is a tool that creates a minimal buildroot and add the listed BuildRequires.  I have built the created srpm on Fedora 10 for ppc, ppc64, i386 and x86_64.  ive not tested the ppc versions but the i386 and x86_64 build don't run. 

5.1.5  doesn't build due to the newly included external components.
[26 Jan 2009 15:10] Susanne Ebrecht
Our development team is doing a lot to get Workbench on Linux working stable. Which also means that they fixed a lot for version 5.1.5.

Please try Workbench 5.1.5 and let us know if this version still have the issue.
[26 Jan 2009 17:49] Dennis Gilmore
fails to build due to the included ctemplate failing to build.  it si really in poor taste to include external projects in your tarball.  please stop that practice and use the host provided ones.  with mysql-gui-tools i have had no end of pain keeping ancient external projects patched to build with newer gcc's 

Fedora will soon have gcc 4.4.0 im honestly very scared that  none of the gui tools will continue to build
[26 Jan 2009 22:10] Dennis Gilmore
after telling the build not to compile ctemplate trying to run i get 

terminate called after throwing an instance of 'grt::grt_runtime_error'
  what():  Type mismatch: expected object of type app.Plugin, but got non-object type
 then the splashscreen sits there and i have to kill mysql-workbench-bin
[26 Jan 2009 22:13] Dennis Gilmore
after fixing /usr/bin/mysql-workbench to point at the correct locations i get 

* Message: Trying to load module '/usr/lib64/mysql-workbench/modules/forms.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/wb.mysql.validation.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/wb.mysql.import.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/db.mysql.sqlparser.grt.so' (cpp)

** (mysql-workbench-bin:14661): WARNING **: Native C++ module classes must have the suffix Impl to avoid confusion between implementation and wrapper classes (MysqlSqlFacade)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/dbutils.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/db.mysql.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/wb.model.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/modules/wb.validation.grt.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/plugins/db.wbp.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/plugins/db.mysql.diff.reporting.wbp.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/plugins/db.mysql.wbp.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/plugins/wb.model.editors.wbp.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/plugins/db.mysql.editors.wbp.so' (cpp)
** Message: Trying to load module '/usr/lib64/mysql-workbench/plugins/wb.model.reporting.wbp.so' (cpp)
** Message: WARNING: Could not get pointer to grt_module_init in module /usr/lib64/mysql-workbench/plugins/wb.model.reporting.wbp.so (`grt_module_init': /usr/lib64/mysql-workbench/plugins/wb.model.reporting.wbp.so: undefined symbol: grt_module_init)
** Message: WARNING: Could not load wb.model.reporting.wbp.so: Invalid module /usr/lib64/mysql-workbench/plugins/wb.model.reporting.wbp.so
[29 Jan 2009 12:38] Susanne Ebrecht
Many thanks for your patient.

We found the reason for your crash. Finally we could reproduce the crashes and could do a deeper analysis of it.

The problem is related to gcc and hit all 64bit systems.

I opened a new bug report where the problem is described detailed.

Because the exact problem is written there I will set this as duplicate of the other bug report.

This is a duplicate of bug #42435