gdc segfaults when compiling gcc/deh.d, how would
I get a stacktrace or some more usable debug info ?

(i.e. the brand new GDC compiler based on GCC 3.4.3,
when bootstrapping the Phobos library in the make)

--anders

PS. Target was "ppc-yellowdog-linux"

Anders F Björklund wrote:
> gdc segfaults when compiling gcc/deh.d, how would
> I get a stacktrace or some more usable debug info ?
> 
> (i.e. the brand new GDC compiler based on GCC 3.4.3,
> when bootstrapping the Phobos library in the make)
> 
> --anders
> 
> PS. Target was "ppc-yellowdog-linux"

Start the "cc1d" executable in gdb, wait for it crash, and type "bt". The data from "info args" and "info locals" would also be useful.

To get the exact command line arguments for cc1d, add the -v option to the gdc command as printed by make with the -v option.

David

David Friedman wrote:

>> gdc segfaults when compiling gcc/deh.d, how would
>> I get a stacktrace or some more usable debug info ?

> Start the "cc1d" executable in gdb, wait for it crash, and type "bt". The data from "info args" and "info locals" would also be useful.

It still crashes with GDC 0.15 and GCC 3.4.4, so I
thought I'd finally get around to doing the trace...

How do I do the "info" commands ? Are those in GDB ?
(as it just said: "No symbol table info available.")

--anders

Anders F Björklund wrote:
> David Friedman wrote:
> 
>>> gdc segfaults when compiling gcc/deh.d, how would
>>> I get a stacktrace or some more usable debug info ?
> 
> 
>> Start the "cc1d" executable in gdb, wait for it crash, and type "bt". The data from "info args" and "info locals" would also be useful.
> 
> 
> It still crashes with GDC 0.15 and GCC 3.4.4, so I
> thought I'd finally get around to doing the trace...
> 
> How do I do the "info" commands ? Are those in GDB ?
> (as it just said: "No symbol table info available.")
> 
> --anders
> 
> 

"info args" and "info locals" should work.  I'm not sure what the error means since there is obviously some debugging information present.  It may be an older gdb/newer compiler issue.  In any case, I think I can figure out the problem from the backtrace.  Thanks!

David