[CMake] Assembler handling in 2.8.5 vs 2.8.4

Alexander Neundorf a.neundorf-work at gmx.net
Tue Aug 16 15:46:50 EDT 2011


On Monday 08 August 2011, Florian Reinhard wrote:
> 2011/8/1 Alexander Neundorf <a.neundorf-work at gmx.net>:
> >> I attached the gdb output i got. I've been running gdb from a minGW32
> >> installation on a normal windows cmd.exe on windows7 x64.
> >> Unfortunately there's no backtrace (or i'm just don't know gdb well
> >> enough) and as soon as i execute "strace" after the cmake crash, gdb
> >> crashes as well.
> > 
> > Can you try to "catch throw" in gdb, then gdb should break when the
> > exception is thrown, and you should be able to get a backtrace then, so
> > we can see where it is thrown.
> 
> Catch throw isn't able to produce a backtrace either. See attachment
> 
> >> I'm compiling a complete firmware for a Texas
> >> Instruments DSP (674x family) and no single binaries.
> > 
> > Just to make sure I understand correctly: the firmware consists of one
> > big binary which contains everything and there are no shared libraries
> > or anything like this, right ?
> 
> Right, everything get's linked together into one single big binary, no
> shared libraries.
> I doubt anyone runs a real OS on these chips. There is the TI OMAP
> platform which have a ARM Core which run Linux, but i can't tell which
> compiler they use for the ARM Core.
> 
> 
> Besides the issues above: the commandline options you provided do work
> with the 7.2.x series but aren't compatible with the 6.1.x series
> which is still maintained by TI but doesn't feature verbose (=long)
> commandline options. Using short commandline options 7.2.x and 6.1.x
> are compatible on CLI.

Probably we should then use the short ones.

> What's the way to go here? Duplicate the code in the .cmake modules,
> keep the more readable 7.2.x version and add the same ones for 6.1.x
> again?

Can you please create a ticket for this in the cmake bug tracker at 
http://public.kitware.com/Bug , and enter this information there ?
Otherwise I'll loose track of it.

Thanks
Alex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.cmake.org/pipermail/cmake/attachments/20110816/6c291c30/attachment.htm>


More information about the CMake mailing list