"global opt framework" entry in MAINTAINERS

"global opt framework" entry in MAINTAINERS

[Andrew Pinski]

Hi Jeff (and others),
  I was looking into the MAINTAINERS file and noticed "global opt
framework" as being maintained by you. This was the first time I had
saw this even though it has been in MAINTAINERS since 1998/9.
I had to go and search and it seems to mean the cfg/basic block
infrastructure and the bitmap.{cc,h}[1]. But I could still be wrong.
Maybe it makes sense to clarify what is meant by "global opt framework".

Note I am not trying to reduce the range of "global opt framework"
areas but rather better understand what it means these days. This is
more for new folks to understand who to CC. It also will help with
labels on the forge and new folks easier who can approve which areas.

As an aside, I think we should try to clean up MAINTAINERS for folks
who are no longer active and/or don't want to contribute any more.
Especially with old emails listed.

Thanks,
Andrew Pinski

[1] https://gcc.gnu.org/news/gcse.html

Re: "global opt framework" entry in MAINTAINERS

[Richard Biener]

On Tue, Jan 20, 2026 at 5:57 AM Andrew Pinski via Gcc <gcc@gcc.gnu.org> wrote:
>
> Hi Jeff (and others),
>   I was looking into the MAINTAINERS file and noticed "global opt
> framework" as being maintained by you. This was the first time I had
> saw this even though it has been in MAINTAINERS since 1998/9.
> I had to go and search and it seems to mean the cfg/basic block
> infrastructure and the bitmap.{cc,h}[1]. But I could still be wrong.
> Maybe it makes sense to clarify what is meant by "global opt framework".
>
> Note I am not trying to reduce the range of "global opt framework"
> areas but rather better understand what it means these days. This is
> more for new folks to understand who to CC. It also will help with
> labels on the forge and new folks easier who can approve which areas.

sourcebuild.texi might be a good place to put more elaborate documentation
on what the maintainer areas actually cover.

> As an aside, I think we should try to clean up MAINTAINERS for folks
> who are no longer active and/or don't want to contribute any more.
> Especially with old emails listed.

There's been issues in the past for people returning and finding themselves
removed.  We should set to document some policy around this.  IMO if
listed people are inactive but upon request say they are still curious about
changes when CCed it's OK to keep them in.

Richard.

>
> Thanks,
> Andrew Pinski
>
> [1] https://gcc.gnu.org/news/gcse.html

Re: "global opt framework" entry in MAINTAINERS

[Jonathan Wakely]

On Tue, 20 Jan 2026 at 04:57, Andrew Pinski via Gcc <gcc@gcc.gnu.org> wrote:
>
> Hi Jeff (and others),
>   I was looking into the MAINTAINERS file and noticed "global opt
> framework" as being maintained by you. This was the first time I had
> saw this even though it has been in MAINTAINERS since 1998/9.
> I had to go and search and it seems to mean the cfg/basic block
> infrastructure and the bitmap.{cc,h}[1]. But I could still be wrong.
> Maybe it makes sense to clarify what is meant by "global opt framework".

FWIW when I saw the email subject (and before I read the paragraph
above) I assumed this was about the .opt files and how driver options
are generated from those files!

Re: "global opt framework" entry in MAINTAINERS

[Mark Wielaard]

Hi,

On Tue, Jan 20, 2026 at 09:49:34AM +0100, Richard Biener via Gcc wrote:
> On Tue, Jan 20, 2026 at 5:57 AM Andrew Pinski via Gcc <gcc@gcc.gnu.org> wrote:
> > As an aside, I think we should try to clean up MAINTAINERS for folks
> > who are no longer active and/or don't want to contribute any more.
> > Especially with old emails listed.
> 
> There's been issues in the past for people returning and finding themselves
> removed.  We should set to document some policy around this.  IMO if
> listed people are inactive but upon request say they are still curious about
> changes when CCed it's OK to keep them in.

We now have a process that disables accounts when not used for more
than a year. It would make sense to have an Emeritus section in
MAINTAINERS listing those accounts. So they are still listed, but
there is a hint that they might not be active at the moment. And if
they wish to become active again it is simple to reactivate their
account: https://sourceware.org/sourceware/accountinfo.html
So they can move their entry from Emeritus to Developer again.

Cheers,

Mark

Re: "global opt framework" entry in MAINTAINERS

[Jakub Jelinek]

On Tue, Jan 20, 2026 at 01:10:52PM +0100, Mark Wielaard wrote:
> On Tue, Jan 20, 2026 at 09:49:34AM +0100, Richard Biener via Gcc wrote:
> > On Tue, Jan 20, 2026 at 5:57 AM Andrew Pinski via Gcc <gcc@gcc.gnu.org> wrote:
> > > As an aside, I think we should try to clean up MAINTAINERS for folks
> > > who are no longer active and/or don't want to contribute any more.
> > > Especially with old emails listed.
> > 
> > There's been issues in the past for people returning and finding themselves
> > removed.  We should set to document some policy around this.  IMO if
> > listed people are inactive but upon request say they are still curious about
> > changes when CCed it's OK to keep them in.
> 
> We now have a process that disables accounts when not used for more
> than a year. It would make sense to have an Emeritus section in
> MAINTAINERS listing those accounts. So they are still listed, but
> there is a hint that they might not be active at the moment. And if
> they wish to become active again it is simple to reactivate their
> account: https://sourceware.org/sourceware/accountinfo.html
> So they can move their entry from Emeritus to Developer again.

Should that be based just on whether the maintainer/reviewer has disabled
accounts?
In theory, a maintainer/reviewer even with no commits within a year might
be active on the mailing list (most importantly in patch review) or in
bugzilla.
If the account is disabled and no mailing list/bugzilla activity (but am not
sure if a year is the right time, if say 3 wouldn't be better), then moving
to Emeritus section looks reasonable.

	Jakub

Re: "global opt framework" entry in MAINTAINERS

[Jeffrey Law]

On 1/19/2026 9:56 PM, Andrew Pinski via Gcc wrote:
> Hi Jeff (and others),
>    I was looking into the MAINTAINERS file and noticed "global opt
> framework" as being maintained by you. This was the first time I had
> saw this even though it has been in MAINTAINERS since 1998/9.
> I had to go and search and it seems to mean the cfg/basic block
> infrastructure and the bitmap.{cc,h}[1]. But I could still be wrong.
> Maybe it makes sense to clarify what is meant by "global opt framework".
>
> Note I am not trying to reduce the range of "global opt framework"
> areas but rather better understand what it means these days. This is
> more for new folks to understand who to CC. It also will help with
> labels on the forge and new folks easier who can approve which areas.
>
> As an aside, I think we should try to clean up MAINTAINERS for folks
> who are no longer active and/or don't want to contribute any more.
> Especially with old emails listed.
I think the intent was to also include the RTL global cprop, pre & 
hoisting bits, but it's been a long time since we put that entry together.

I think the question we need to answer is more a "where/how" to document 
what a given entry means.  I guess we can add clarifying text as its own 
section in the MAINTAINERS file.  Or a separate file.  Doesn't matter 
much to me as  long as it's easy to find.

Jeff