[PATCH] RTEMS: Use strict DWARF-2 on ARM, PowerPC, SPARC

[Gedare Bloom]

On Tue, Jun 18, 2013 at 1:32 PM, Ralf Corsepius
<ralf.corsepius at rtems.org> wrote:
> On 06/18/2013 02:24 PM, Sebastian Huber wrote:
>>
>> On 06/18/2013 02:10 PM, Ralf Corsepius wrote:
>>>
>>> On 06/18/2013 01:58 PM, Sebastian Huber wrote:
>>>>
>>>> Some debuggers do not cope with the new DWARF3/4 debug format introduced
>>>> with GCC 4.8.  Default to strict DWARF-2 on ARM, PowerPC and SPARC for
>>>> now.
>>>>
>>>> This patch should be committed to GCC 4.8 and 4.9.
>>>
>>>
>>> I am opposed to this patch, because
>>>
>>> * GNU software should not care about the limitations of commerical
>>> stuff and
>>> should only care about gdb.
>>
>>
>> Actually GCC cares about commercial stuff, e.g. the VxWorks and Darwin
>> ports use exactly the same mechanism.
>
> Well, RTMES these OSes are _closed source_. If GCC is interested in playing
> it nice to them (Which I consider to be a management fault), they have no
> other choice.
>
> RTEMS is open source, therefore there is no need to "play it nice" to any
> closed source debugger stuff or similar.
>
Unless RTEMS developers particularly like to use the closed source
debugger. Anyway, I think someone mentioned the sparc GDB has a
problem too.

>
>> The SPARC version of GDB seems to have problems here also:
>>
>> http://www.rtems.org/pipermail/rtems-devel/2013-May/003188.html
>>
>>>
>>> * We should stay with the GCC's defaults and not diverge from these.
>>
>>
>> In general this is true.
>>
>>> * Users, who are facing issues with commerical stuff can always
>>> manually pass
>>> appropriate options to CC if they need it.
>>
>>
>> This approach is not possible for the multilibs.
>
>
> It surly is. Simply let alone and forget about these commercial tools and
> help improving the GNU versions.
>
> People who are using these commercial tools need to comprehend, that they
> are on their own and need to cope with the problems *their* choice implies
> are homemade by themselves.
>
Actually, my experience has been that relying on open source tools
often leaves me on my own coping with problems.

> In other words, if GCC's debug format doesn't work with the tools you spent
> a lot of money for, it's your privat mistake to have spent money on them.
>
> It definitely is not a valid reason to strangle RTEMS!
>
Such colorful language. I don't see how this patch poses a problem for
anyone right now. Someday perhaps the DWARF3/4 support will be mature
and usable in production environments, but today does not seem to be
that day.

Persisting in attacks on this patch here and on the gcc-patches ml
because of closed-mindedness about what qualifies as "open source"
would be silly childlike behavior.

>
> Ralf
>
>
> _______________________________________________
> rtems-devel mailing list
> rtems-devel at rtems.org
> http://www.rtems.org/mailman/listinfo/rtems-devel