Re: GNU Make Errors in Windows Application Event Log

[B G Ramprasad Torati]

Hi  JG,

Thanks for the response.


On Sun, 31 Oct 2004 21:53:30 +0000, J. Grant <address@hidden> wrote:
> Hi Ramp,
> 
> > I'm using theGNU Make version 3.79.1 Built for Windows2000
> 
> You may like to try version 3.80, please add hash.obj to NMakefile if
> building using MS nmake.exe.

The problem is there with 3.80 as well. I am not building with
nmake.exe, instead I am using the 'build_w32.bat' provided with the
make source, which uses cl.exe and link.exe.

> 
> 
> 
> 
> > MultiProcessor Server and the following Windows Error event occurs:
> >
> >  EventID:   0
> >  Category:  0
> >  Unhandled exception filter called from program make
> >  ExceptionCode = c0000005
> >  ExceptionFlags = 0
> >  ExceptionAddress = 4166e4
> >  Access violation: read operation at address 3f
> 
> Could be a bug in your compiler.
> 
> Just to confirm, is this your gnumake.exe which has this exeption? Or
> one of the processes which gnumake.exe launches?

I can confirm that this exception is from gnumake.exe only, which must
be triggered by one of the processes gnumake.exe launches. The debug
trace showed that the exception was indeed coming from 'process_easy'
call of 'exec_command' defined in 'job.c' file, which actually
launches the command process.

> 
> Can you try with the debug build of gnumake.exe for win32
> 
> > I see this error event (there are 250 of them) only when try to build
> > on a fast multiprocessor build machine. Not with a single p4 based
> > machine. Surprisingly this doesn't stop the build process and
> > everything gets built to the end. I've tried to debug the problem
> > through VC6, but no luck yet.
> 
> Can you set "first-chance exceptions" to all break into the debugger.

I am using the debug build of gnumake.exe to debug the problem. How to
set 'fist-chance exceptions'. Please add full instructions for setting
(and for restoring!) the "first-chance exceptions".

> 
> > Since the stuff gets built to the end, What I would like to know,  is
> > that these Event Log Entries really are benign and are not going to
> > wreck any of my future build. Can somebody help in getting this
> > assurance?
> 
> Can you provide some more logs from your compile and also a testcase to
> replicate this problem.
> 
Unfortunately I couldn't create this problem with a slow machine or a
simple test case. I am attaching the imported windows event long from
the failures reported machine. The build takes about 7 hours on a fast
machine and involves in building about 7000+ components (i.e. atleast
those many make.exe invocations), but reporting only around 250
exceptions with make.exe. That means not all the make.exe calls
reporting the problem.

Thanks and Regards,
Ramp

> Kind regards
> 
> JG
> 
> --
> Homepage: http://jguk.org/
> Blog: http://jguk.org/#blog
> Radio: http://jguk.org/#radio
> 
> 


-- 
Cheers..
Ramp

AppEvents[1].csv
Description: Binary data