[PATCH] rtems-binutils-head.cfg, rtems-gdb-head.cfg: Bump to 3b2bef8

[Chris Johns]

On 28/5/21 6:01 am, Joel Sherrill wrote:
> On Wed, May 26, 2021 at 2:29 AM Chris Johns <chrisj at rtems.org
> <mailto:chrisj at rtems.org>> wrote:
> 
>     I am ok with this going in but ....
> 
>     I am becoming concerned we are building a debt around MacOS and Windows in the
>     tools that we need to resolve. Our ability to release when we would like to
>     depends on the state of the tools and the state of the tools for all archs. :)
> 
> 
> I don't disagree. This bump was actually necessary because gdb wouldn't
> build on CentOS 7. It had a bug and didn't add -std=c99 and was using 
> C99 code.

If this position was taken for Windows or some archs would we have moved to GDB
head like we have? Does Linux by itself define a stable build suitable for a
change? I do not think it does. If this move fixed Linux but broke Windows is
making the change valid?

I suggest it is in the interest of everyone to see the tools fixed on all hosts.
I am not sure how we stage changes like this so they can be tested as requiring
each developer to have access to all hosts is not feasible.

I have a real concern that fixing some of these issues may require regressing
tool changes and with projects like pre-qual there may be a negative impact.

> Crossing threads, the mailer series of modifications by Alex was to support my
> need to send reports from Windows. The addition of the various smtp options
> and using gitconfig for email options. My intent was to turn on some Windows
> testing (Mingw and Cygwin) but I couldn't get test reports to build at . I'm not 
> tackling MacOS but was trying to get reports on the others which is the first
> step.

This is a great addition and a good solution. It however is an after the fact
change and I am wondering how we pre-test changes?

> I know that's not apparent from the changes themselves but that was the
> overall goal.

I knew it was and understood this.

> Also, I think the mailer changes were already in either RSB or rtems-tools 
> and needed to be synced. If something was missed when the patches were
> reviewed and committed in the first repo, that can be addressed in a follow up
> patch.
> 
> I'm trying to increase my testing from 3 hosts to 5 and need reports from them
> all to show up on build at .

Thank you for your efforts. They make a huge difference in the quality we can
produce.

Chris