Source builder mingw32 fails in gdb

[Chris Johns]

On 29/10/2013 9:43 am, gds wrote:
> On 10/28/2013 04:32 PM, Chris Johns wrote:
>>
>> Does 4.11 for MinGW32 build ok ?
>
> Once I figured out how to go around the firewall here and do the cvs
> checkout on "Sept-2013" newlib (set env. var. CVSROOT) I thought it was
> going to work. However, my old version of mingw32 on fedora f14 seg
> faults when compiling gcc/var-tracking.c:
>
> ../../gcc-4.8.1/gcc/var-tracking.c: In function 'void
> add_stores(rtx_def*, const rtx_def*, void*)':
> ../../gcc-4.8.1/gcc/var-tracking.c:5782:1: internal compiler error:
> Segmentation fault

Ouch.

>
> $ i686-pc-mingw32-g++ --version
> i686-pc-mingw32-g++ (GCC) 4.5.0 20100414 (Fedora MinGW 4.5.0-1.fc14)
> (Yes, it is using g++ to compile this.)
>
> I will try this on the f16 version of i686-pc-mingw32-g++ later.
>

Ok.

> One other thing that I've seen is that you have to have python-devel
> installed. Without this, the gdb build also fails. This isn't found by
> the source-builder/sb-check.

Yes it is not checked.

> It appears all "gdb configure" is looking for is Python.h but the error
> it prints sound like you don't have python installed, or that you have
> the wrong version.

I am building gdb with Python so we make sure Python and GDB does work, 
plus it is useful when working with RTEMS. We have some support for 
RTEMS in GDB Python scripts and I would this support to improve. I do 
not wish to turn off this support.

It would be nice to have the RSB check for these things however I am not 
sure how I would do this without becoming closely coupled with 
particular packaging systems.

Chris