mcore-*-* target to be deleted?

mcore-*-* target to be deleted?

[Brian LaPonsey]

Hello,

I (belatedly) noticed an announcement on the GDB web page that the mcore
target has been declared obsolete and is going to be deleted from gdb in
the next release.

I have made some inquiries about the impact this would have on our
customers, and although the response wasn't huge, it was definitely not
zero.  There's a small but determined population of developers who
continue to use gdb with mcore.  It also has a small but growing market
with small firms working through distributors, as well as a new design
group in China that prefers to use non-Microsoft tools.

I don't know if one vote matters, but I vote to save the mcore.

Thanks,
Brian

Re: mcore-*-* target to be deleted?

[Peter Barada]

>I (belatedly) noticed an announcement on the GDB web page that the mcore
>target has been declared obsolete and is going to be deleted from gdb in
>the next release.
>
>I have made some inquiries about the impact this would have on our
>customers, and although the response wasn't huge, it was definitely not
>zero.  There's a small but determined population of developers who
>continue to use gdb with mcore.  It also has a small but growing market
>with small firms working through distributors, as well as a new design
>group in China that prefers to use non-Microsoft tools.
>
>I don't know if one vote matters, but I vote to save the mcore.

If you know of(or can support) someone who agrees to be a maintainer
for the target then you'll have a *much* better chance of pulling
mcore off of the 'obsolete' list.

You'll also want to look to see the support status of mcore in GCC and
Newlib...

-- 
Peter Barada
peter@the-baradas.com

Re: mcore-*-* target to be deleted?

[Daniel Jacobowitz]

On Thu, Dec 02, 2004 at 02:04:01PM +0000, Brian LaPonsey wrote:
> Hello,
> 
> I (belatedly) noticed an announcement on the GDB web page that the mcore
> target has been declared obsolete and is going to be deleted from gdb in
> the next release.
> 
> I have made some inquiries about the impact this would have on our
> customers, and although the response wasn't huge, it was definitely not
> zero.  There's a small but determined population of developers who
> continue to use gdb with mcore.  It also has a small but growing market
> with small firms working through distributors, as well as a new design
> group in China that prefers to use non-Microsoft tools.
> 
> I don't know if one vote matters, but I vote to save the mcore.

A vote doesn't make much difference - someone willing to maintain the
code, update it, and regularly test it, would make more of a difference.
The target has already been removed in HEAD, and we won't re-add it
unless a maintainer steps forward and submits an updated port.

-- 
Daniel Jacobowitz

Re: mcore-*-* target to be deleted?

[Andrew Cagney]

Peter Barada wrote:
>>I (belatedly) noticed an announcement on the GDB web page that the mcore
>>target has been declared obsolete and is going to be deleted from gdb in
>>the next release.
>>
>>I have made some inquiries about the impact this would have on our
>>customers, and although the response wasn't huge, it was definitely not
>>zero.  There's a small but determined population of developers who
>>continue to use gdb with mcore.  It also has a small but growing market
>>with small firms working through distributors, as well as a new design
>>group in China that prefers to use non-Microsoft tools.
>>
>>I don't know if one vote matters, but I vote to save the mcore.

The mcore code has, unfortunatly, become very out of date.   The 
mainline hasn't built since 6.2 was released, and with 6.3 past that 
broken code was dropped.  This all came about because GDB's frame code 
was rewritten from scratch and the mcore was never updated to work with 
the new code.

The way to revive the mcore isn't through votes, nominations, or 
appointments, it's through code.  A developer with an assignment 
contributes code to current standards (most importantly does not use 
deprecated interfaces) then it can very easily be re-integrated (and 
they may quickly find themselves in a more responsible role).  Any 
questions on how to do it welcome(1).

Andrew

(1) Exprerience suggests that largely starting from scratch (using the 
existing code as a guideline) is easier.  This is because the bulk of 
the current file is redundant.