Nominations for Anthropic’s Claude for Open Source program

Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

Developers,

Anthropic is generously offering 6 months of Claude Max 20x to Free and Open
Source Software projects to support the FOSS ecosystem. These tools can be
incredibly helpful with reviewing upstream code, evaluating the conformance of
a particular interface to a standard, or just for your own learning and
exploring the implementation of the project or subsystem.

If you are an established, active contributor to the GNU Toolchain and are
interested in having access to Claude Max 20x under Anthropic’s Claude for Open
Source program, please respond to this email thread with a short description of
how you would utilize the access and the potential impact on the GNU Toolchain.
The program's slots are limited and David Edelsohn and I plan to review the
proposals. Please keep in mind that these are individual grants from Anthropic
to you personally. If you are approved you’ll get an email from Anthropic with
further instructions.

While the current GNU Project policy is not to accept LLM-generated content in
the projects, that still leaves other uses like bug triage, patch review, and
research as ways to use these tools.

--
Cheers,
Carlos.

P.S. This is not an endorsement of Anthropic, Claude or LLMs. We dislike
having to say it, but please don’t use this thread as a platform to express
your personal opinions about the companies or about AI’s impact on the world.

Re: Nominations for Anthropic’s Claude for Open Source program

[Andrew Pinski via Gdb]

On Fri, Jun 12, 2026, 12:19 PM Carlos O'Donell <carlos@redhat.com> wrote:

> Developers,
>
> Anthropic is generously offering 6 months of Claude Max 20x to Free and
> Open
> Source Software projects to support the FOSS ecosystem. These tools can be
> incredibly helpful with reviewing upstream code, evaluating the
> conformance of
> a particular interface to a standard, or just for your own learning and
> exploring the implementation of the project or subsystem.
>
> If you are an established, active contributor to the GNU Toolchain and are
> interested in having access to Claude Max 20x under Anthropic’s Claude for
> Open
> Source program, please respond to this email thread with a short
> description of
> how you would utilize the access and the potential impact on the GNU
> Toolchain.
> The program's slots are limited and David Edelsohn and I plan to review the
> proposals. Please keep in mind that these are individual grants from
> Anthropic
> to you personally. If you are approved you’ll get an email from Anthropic
> with
> further instructions.
>
> While the current GNU Project policy is not to accept LLM-generated
> content in
> the projects, that still leaves other uses like bug triage, patch review,
> and
> research as ways to use these tools.
>
> --
> Cheers,
> Carlos.
>
> P.S. This is not an endorsement of Anthropic, Claude or LLMs. We dislike
> having to say it, but please don’t use this thread as a platform to express
> your personal opinions about the companies or about AI’s impact on the
> world.
>

Dont send out emails like this in the first place then.

Also note coverity has been offering similar services for code coverage and
static analysis for years now. But there was no notice like that sent out
for this. Why try to push llm when static analysis does the same?




>

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/12/26 3:29 PM, Andrew Pinski wrote:
> Dont send out emails like this in the first place then.
> 
> Also note coverity has been offering similar services for code
> coverage and static analysis for years now. But there was no notice
> like that sent out for this. Why try to push llm when static
> analysis does the same?

Coverity FOSS scanning of glibc has been in the release process
for almost 11 years [1][2][3].

Emails to this list are permitted so long as they are on topic,
and meet the community code of conduct.

Asking politely to stay on topic is appropriate.

Telling the community what it should or should not do is not
called for.

-- 
Cheers,
Carlos.

[1] https://inbox.sourceware.org/libc-alpha/443E97C6.2080801@coverity.com/
[2] https://inbox.sourceware.org/libc-alpha/55DFBA26.2000601@redhat.com/
[3] https://sourceware.org/glibc/wiki/Coverity

Re: Nominations for Anthropic’s Claude for Open Source program

[Andrew Pinski via Gdb]

On Fri, Jun 12, 2026 at 2:18 PM Carlos O'Donell <carlos@redhat.com> wrote:
>
> On 6/12/26 3:29 PM, Andrew Pinski wrote:
> > Dont send out emails like this in the first place then.
> >
> > Also note coverity has been offering similar services for code
> > coverage and static analysis for years now. But there was no notice
> > like that sent out for this. Why try to push llm when static
> > analysis does the same?
>
> Coverity FOSS scanning of glibc has been in the release process
> for almost 11 years [1][2][3].
>
> Emails to this list are permitted so long as they are on topic,
> and meet the community code of conduct.
>
> Asking politely to stay on topic is appropriate.
>
> Telling the community what it should or should not do is not
> called for.

And your email is telling the community it should be NOT talking about
the ethics here.
So you just violated your own writing.

>
> --
> Cheers,
> Carlos.
>
> [1] https://inbox.sourceware.org/libc-alpha/443E97C6.2080801@coverity.com/
> [2] https://inbox.sourceware.org/libc-alpha/55DFBA26.2000601@redhat.com/
> [3] https://sourceware.org/glibc/wiki/Coverity
>

Re: Nominations for Anthropic’s Claude for Open Source program

[Adhemerval Zanella Netto via Gdb]

On 12/06/26 16:29, Andrew Pinski wrote:
> 
> 
> On Fri, Jun 12, 2026, 12:19 PM Carlos O'Donell <carlos@redhat.com <mailto:carlos@redhat.com>> wrote:
> 
>     Developers,
> 
>     Anthropic is generously offering 6 months of Claude Max 20x to Free and Open
>     Source Software projects to support the FOSS ecosystem. These tools can be
>     incredibly helpful with reviewing upstream code, evaluating the conformance of
>     a particular interface to a standard, or just for your own learning and
>     exploring the implementation of the project or subsystem.
> 
>     If you are an established, active contributor to the GNU Toolchain and are
>     interested in having access to Claude Max 20x under Anthropic’s Claude for Open
>     Source program, please respond to this email thread with a short description of
>     how you would utilize the access and the potential impact on the GNU Toolchain.
>     The program's slots are limited and David Edelsohn and I plan to review the
>     proposals. Please keep in mind that these are individual grants from Anthropic
>     to you personally. If you are approved you’ll get an email from Anthropic with
>     further instructions.
> 
>     While the current GNU Project policy is not to accept LLM-generated content in
>     the projects, that still leaves other uses like bug triage, patch review, and
>     research as ways to use these tools.
> 
>     --
>     Cheers,
>     Carlos.
> 
>     P.S. This is not an endorsement of Anthropic, Claude or LLMs. We dislike
>     having to say it, but please don’t use this thread as a platform to express
>     your personal opinions about the companies or about AI’s impact on the world.
> 
> 
> Dont send out emails like this in the first place then.
> 
> Also note coverity has been offering similar services for code coverage and static analysis for years now. But there was no notice like that sent out for this. Why try to push llm when static analysis does the same?


gnulib guys seem to be experimenting with LLM for patch review recently [1],
and it has uncovered real regressions [2]. And I recall that their 
experience with Coverity has been subpar [3], and I think they even started 
disregarding it.

Personally, I see value in adding an LLM to the patch review phase, especially 
since we lack manpower and initial triage can help newcomers and uncover some 
issues.


[1] https://lists.gnu.org/archive/html/bug-gnulib/2026-06/msg00016.html
[2] https://lists.gnu.org/archive/html/bug-gnulib/2026-06/msg00037.html
[3] https://lists.gnu.org/archive/html/bug-gnulib/2019-05/msg00059.html

Re: Nominations for Anthropic’s Claude for Open Source program

[Andrew Pinski via Gdb]

On Fri, Jun 12, 2026, 4:07 PM Adhemerval Zanella Netto <
adhemerval.zanella@linaro.org> wrote:

>
>
> On 12/06/26 16:29, Andrew Pinski wrote:
> >
> >
> > On Fri, Jun 12, 2026, 12:19 PM Carlos O'Donell <carlos@redhat.com
> <mailto:carlos@redhat.com>> wrote:
> >
> >     Developers,
> >
> >     Anthropic is generously offering 6 months of Claude Max 20x to Free
> and Open
> >     Source Software projects to support the FOSS ecosystem. These tools
> can be
> >     incredibly helpful with reviewing upstream code, evaluating the
> conformance of
> >     a particular interface to a standard, or just for your own learning
> and
> >     exploring the implementation of the project or subsystem.
> >
> >     If you are an established, active contributor to the GNU Toolchain
> and are
> >     interested in having access to Claude Max 20x under Anthropic’s
> Claude for Open
> >     Source program, please respond to this email thread with a short
> description of
> >     how you would utilize the access and the potential impact on the GNU
> Toolchain.
> >     The program's slots are limited and David Edelsohn and I plan to
> review the
> >     proposals. Please keep in mind that these are individual grants from
> Anthropic
> >     to you personally. If you are approved you’ll get an email from
> Anthropic with
> >     further instructions.
> >
> >     While the current GNU Project policy is not to accept LLM-generated
> content in
> >     the projects, that still leaves other uses like bug triage, patch
> review, and
> >     research as ways to use these tools.
> >
> >     --
> >     Cheers,
> >     Carlos.
> >
> >     P.S. This is not an endorsement of Anthropic, Claude or LLMs. We
> dislike
> >     having to say it, but please don’t use this thread as a platform to
> express
> >     your personal opinions about the companies or about AI’s impact on
> the world.
> >
> >
> > Dont send out emails like this in the first place then.
> >
> > Also note coverity has been offering similar services for code coverage
> and static analysis for years now. But there was no notice like that sent
> out for this. Why try to push llm when static analysis does the same?
>
>
> gnulib guys seem to be experimenting with LLM for patch review recently
> [1],
> and it has uncovered real regressions [2]. And I recall that their
> experience with Coverity has been subpar [3], and I think they even
> started
> disregarding it.
>
> Personally, I see value in adding an LLM to the patch review phase,
> especially
> since we lack manpower and initial triage can help newcomers and uncover
> some
> issues.
>


All of this talk about llm and llm reviews is off-topic for this list.
Including the original email is off-topic and should not have been sent.

Carlos,
  Can you step down from the llm gcc working group for this email since it
shows a conflict of interest of prompting one product.  Plus it shows poor
taste of sending out an email about a non free product to a mailing list
about glibc development.
Can I also ask you to step down from CoC for violating the CoC here?

Thanks,
Andrea



>
>
>
> [1] https://lists.gnu.org/archive/html/bug-gnulib/2026-06/msg00016.html
> [2] https://lists.gnu.org/archive/html/bug-gnulib/2026-06/msg00037.html
> [3] https://lists.gnu.org/archive/html/bug-gnulib/2019-05/msg00059.html
>

Re: Nominations for Anthropic’s Claude for Open Source program

[Mark Wielaard]

Hi,

On Fri, 2026-06-12 at 16:11 -0700, Andrew Pinski via Gdb wrote:
> All of this talk about llm and llm reviews is off-topic for this list.
> Including the original email is off-topic and should not have been sent.

Agreed, promoting the use of a commercial, non-free, proprietary llm
through a saas model is not appropriate and off-topic for the GNU
development lists.

> Carlos,
>   Can you step down from the llm gcc working group for this email since it
> shows a conflict of interest of prompting one product.  Plus it shows poor
> taste of sending out an email about a non free product to a mailing list
> about glibc development.
> Can I also ask you to step down from CoC for violating the CoC here?

Yeah. It is hard to not see a pattern here. But not sure it is enough
for banning someone. And we all make mistakes from time to time. But we
do want to avoid the appearance of a conflict of interest.

Cheers,

Mark

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/15/26 7:39 AM, Mark Wielaard wrote:
> Hi,
> 
> On Fri, 2026-06-12 at 16:11 -0700, Andrew Pinski via Gdb wrote:
>> All of this talk about llm and llm reviews is off-topic for this list.
>> Including the original email is off-topic and should not have been sent.
> 
> Agreed, promoting the use of a commercial, non-free, proprietary llm
> through a saas model is not appropriate and off-topic for the GNU
> development lists.

I disagree.

For all the reasons David lists here and more:
https://inbox.sourceware.org/gcc/CAGWvnyk1x47jPWs3ML01J-TrgKt8ByALj3HzPpZodVS_73VSBg@mail.gmail.com/

We should always be able to *talk* about non-free alternatives, how to use
them, what they do, how to create alternatives, and how to move away from them.

We don't want to require anyone to use these tools and we don't.

>> Carlos,
>>    Can you step down from the llm gcc working group for this email since it
>> shows a conflict of interest of prompting one product.  Plus it shows poor
>> taste of sending out an email about a non free product to a mailing list
>> about glibc development.
>> Can I also ask you to step down from CoC for violating the CoC here?
> 
> Yeah. It is hard to not see a pattern here. But not sure it is enough
> for banning someone. And we all make mistakes from time to time. But we
> do want to avoid the appearance of a conflict of interest.
Sorry, could you please expand on the pattern you see?

I will raise this with the WG lead Jonathan Wakely this Friday, and ask
the WG's opinion if they believe I can be impartial.

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Mark Wielaard]

Hi Carlos,

On Mon, Jun 15, 2026 at 05:11:18PM -0400, Carlos O'Donell wrote:
> On 6/15/26 7:39 AM, Mark Wielaard wrote:
> >On Fri, 2026-06-12 at 16:11 -0700, Andrew Pinski via Gdb wrote:
> >>All of this talk about llm and llm reviews is off-topic for this list.
> >>Including the original email is off-topic and should not have been sent.
> >
> >Agreed, promoting the use of a commercial, non-free, proprietary llm
> >through a saas model is not appropriate and off-topic for the GNU
> >development lists.
> 
> I disagree.
> 
> For all the reasons David lists here and more:
> https://inbox.sourceware.org/gcc/CAGWvnyk1x47jPWs3ML01J-TrgKt8ByALj3HzPpZodVS_73VSBg@mail.gmail.com/
> 
> We should always be able to *talk* about non-free alternatives, how to use
> them, what they do, how to create alternatives, and how to move away from them.

As explained to David nobody said you cannot talk about alternatives.
But that is something completely different from advertising the
specific use of a commercial, non-free, proprietary llm.

> >>Carlos,
> >>   Can you step down from the llm gcc working group for this email since it
> >>shows a conflict of interest of prompting one product.  Plus it shows poor
> >>taste of sending out an email about a non free product to a mailing list
> >>about glibc development.
> >>Can I also ask you to step down from CoC for violating the CoC here?
> >
> >Yeah. It is hard to not see a pattern here. But not sure it is enough
> >for banning someone. And we all make mistakes from time to time. But we
> >do want to avoid the appearance of a conflict of interest.
> Sorry, could you please expand on the pattern you see?

Just that the David and Carlos show seems to have this pattern of
maximal drama. Where you promote some corporate deal so "generous"
that people shouldn't question whether it is ethical, undermines the
community or the gnu project goals. Then when people obviously do
point out these issues you start attacking the community volunteers
that do the actual work and pick a needless fight with the FSF about
"purity" and "leadership".

Cheers,

Mark

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/16/26 7:03 AM, Mark Wielaard wrote:
> On Mon, Jun 15, 2026 at 05:11:18PM -0400, Carlos O'Donell wrote:
>> On 6/15/26 7:39 AM, Mark Wielaard wrote:
> As explained to David nobody said you cannot talk about alternatives.
> But that is something completely different from advertising the
> specific use of a commercial, non-free, proprietary llm.

I disagree.

There is no hard bright line that distinguishes between what you call
advertising and talking about the options.

My position is to allow all such conversations between anyone on the list.

Where I draw the line, and where it can be drawn is in what we write down
in our documentation, likewise in our process. Such documentation should
reference and require only FOSS tools. There have been several discussions
about this on the gnu-prog mailing list.

>> Sorry, could you please expand on the pattern you see?
> 
> Just that the David and Carlos show seems to have this pattern of
> maximal drama. Where you promote some corporate deal so "generous"
> that people shouldn't question whether it is ethical, undermines the
> community or the gnu project goals. Then when people obviously do
> point out these issues you start attacking the community volunteers
> that do the actual work and pick a needless fight with the FSF about
> "purity" and "leadership".

Thanks for clarifying.

You and anyone else in the community are always free to ask questions.

I disagree with the rest of your characterization of my actions.

The fact that we disagree about certain non-technical direction
doesn't mean we can't work together though in areas where we do agree.

I was in the Sourceware office hours last week, and I plan to continue
to do that to support and coordinate.

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Alexandre Oliva via Gdb]

On Jun 16, 2026, "Carlos O'Donell" <carlos@redhat.com> wrote:

> There is no hard bright line that distinguishes between what you call
> advertising and talking about the options.

I can agree with that statement.  There is room for judgment calls, and
there is room for disagreement about judgment calls.

Now, if it was another mailing list member forwarding spam that
advertised a 30-days-gratis bait for a service entirely misaligned with
the purposes of GNU in general and of GNU libc specifically, to the
point of opposing those purposes, would you still be favorable to it?

Has Anthropic at least made a donation to the GNU Toolchain fund to pay
for this advertisement, or are we such losers that we advertise such
traps free of charge, and even fight over whether polluting our forum
with their spam is appropriate?

Will your stance change when bots subscribe to these lists, and then
start posting advertisement targeted at our developers but otherwise
bearing no relationship with the project?  Will you then call for such
spam to be blocked, or will we all have to endure it because some day
you decreed that it was ok?

> Where I draw the line, and where it can be drawn is in what we write down
> in our documentation, likewise in our process.

Our mailing lists are part of the project as well, and people who take
part in the project or even lead it are expected to behave in line with
the project, rather than against it, at the very least when
participating in project fora and activities.

Advertising services antagonic to and destructive of user freedom is
really way out of line.

-- 
Alexandre Oliva, happy hacker            https://blog.lx.oliva.nom.br/
Free Software Activist     FSFLA co-founder     GNU Toolchain Engineer
Learn the truth about Richard Stallman at https://stallmansupport.org/

Re: Nominations for Anthropic’s Claude for Open Source program

[Eli Zaretskii via Gdb]

> Cc: Mark Wielaard <mark@klomp.org>, Andrew Pinski <pinskia@gmail.com>,
>  Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>, glibc
>  developers <libc-alpha@sourceware.org>, gdb developers <gdb@sourceware.org>
> Date: Wed, 17 Jun 2026 13:28:57 -0300
> From: Alexandre Oliva via Gdb <gdb@sourceware.org>
> 
> Now, if it was another mailing list member forwarding spam that
> advertised a 30-days-gratis bait for a service entirely misaligned with
> the purposes of GNU in general and of GNU libc specifically, to the
> point of opposing those purposes, would you still be favorable to it?

That's a significant exaggeration, to put it mildly.  I invite you
(and everyone else who wants to make up their minds independently) to
re-read the original message posted by David.  I find no advertisement
there, let alone spam.

Here's that message again:

  Anthropic is generously offering 6 months of Claude Max 20x to Free
  and Open Source Software projects to support the FOSS
  ecosystem. These tools can be incredibly helpful with reviewing
  upstream code, evaluating the conformance of a particular interface
  to a standard, or just for your own learning and exploring the
  implementation of the project or subsystem.

  If you are an established, active contributor to the GNU Toolchain
  and are interested in having access to Claude Max 20x under
  Anthropic’s Claude for Open Source program, please respond to this
  email thread with a short description of how you would utilize the
  access and the potential impact on the GNU Toolchain.  The program's
  slots are limited and Carlos O'Donell and I plan to review the
  proposals. Please keep in mind that these are individual grants from
  Anthropic to you personally. If you are approved you’ll get an email
  from Anthropic with further instructions.

  While the current GNU Project policy is not to accept LLM-generated
  content in the projects, that still leaves other uses like bug
  triage, patch review, and research as ways to use these tools.

This basically says that if someone wants to use the 6-month offer to
work on GNU tools, they should describe their plans and apply for a
slot in using the tools.  That's it.  It even explicitly says in a
P.S. that "This is not an endorsement of Anthropic, Claude or LLMs".

Please be fair and don't accuse people who post here in something they
didn't say.  You don't have to agree, and you don't need to volunteer
to use Anthropic's proposal, but let's have minimally-civil
discussions here, okay?  Even your selection of
intentionally-denigrating words ("spam", "losers", etc.) is highly
inappropriate in civilized discussions, given what the original
message actually said.

> Will your stance change when bots subscribe to these lists, and then
> start posting advertisement targeted at our developers but otherwise
> bearing no relationship with the project?  Will you then call for such
> spam to be blocked, or will we all have to endure it because some day
> you decreed that it was ok?

Not relevant.  Blocking spam and bots, regardless of the content of
those messages, is the job of the list moderators and they generally
do their job just fine so far.

> > Where I draw the line, and where it can be drawn is in what we write down
> > in our documentation, likewise in our process.
> 
> Our mailing lists are part of the project as well, and people who take
> part in the project or even lead it are expected to behave in line with
> the project, rather than against it, at the very least when
> participating in project fora and activities.
> 
> Advertising services antagonic to and destructive of user freedom is
> really way out of line.

Again, there was no advertising.  This is uncalled-for.  Please find
more polite words to express your disagreement (if you must: after
all, that was a call for volunteers, and you don't have to volunteer
nor respond if you don't like the proposal).

More generally, I'm disappointed by the hostility and the general
style some people allow themselves to use in this discussion.  We
don't deserve to call ourselves a community that promotes a better
society if we cannot talk between ourselves in a kinder manner.  Shame
on us!

Re: Nominations for Anthropic’s Claude for Open Source program

[Alexandre Oliva via Gdb]

On Jun 18, 2026, Eli Zaretskii <eliz@gnu.org> wrote:

>> Cc: Mark Wielaard <mark@klomp.org>, Andrew Pinski <pinskia@gmail.com>,
>> Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>, glibc
>> developers <libc-alpha@sourceware.org>, gdb developers <gdb@sourceware.org>
>> Date: Wed, 17 Jun 2026 13:28:57 -0300
>> From: Alexandre Oliva via Gdb <gdb@sourceware.org>
>> 
>> Now, if it was another mailing list member forwarding spam that
>> advertised a 30-days-gratis bait for a service entirely misaligned with
>> the purposes of GNU in general and of GNU libc specifically, to the
>> point of opposing those purposes, would you still be favorable to it?

> That's a significant exaggeration

It's not.

The offer is promotional, for a short time, to attract new users.
Indeed, it fits the pattern that companies use with influencers
nowadays, offering them gratis access to a product for them to publish
their use of the product/service and thereby promote the product/service
to an even broader public.  Companies usually pay for that.

Now, maybe you're not disputing that it was advertisement.

Maybe you're disputing that it's misaligned with GNU.

Surely you're aware that the advertised service, being a service rather
than software you can install and run on your computer to do your
computing, is SaaSS, that denies users the freedoms they deserve.

GNU's prime purpose is to give users a certain set of user freedoms that
are essential for users to have control over their computing, so that
users have the control that the service provider intentionally denies.

I know you value those services regardless.  That's irrelevant for the
assessment.  My argument stands solely on the facts that it's
advertisement of a service that denies users the freedoms they deserve.

It would be no worse if I were very fond of Windows or AIX and got an
request from their providers to get them new influential users through a
similar promotional stunt.

Both amount to promoting freedom-denying "generous" bait in a forum that
exists for the primary purpose of giving users those very freedoms.

> P.S. that "This is not an endorsement of Anthropic, Claude or LLMs".

"It's not what you're thinking"

> Not relevant.  Blocking spam and bots, regardless of the content of
> those messages, is the job of the list moderators and they generally
> do their job just fine so far.

They seem to have missed the post that started this thread, and the
other twin thread in GCC and binutils, otherwise we wouldn't be having
this discussion.

> you don't have to volunteer
> nor respond if you don't like the proposal).

You seem to have missed my response in which I volunteered.  Because the
offer is so harmful that I feel a moral duty to take up a slot so that
it doesn't end up harming anyone.

> if we cannot talk between ourselves in a kinder manner

Being kind encompasses not extending dangerous and harmful offers onto
your colleagues.  Smoke your crack if you must, but please don't try to
"help" others get hooked through "generous" offers in here.  It's not
kind.  It's not cool.  It's detrimental.  It's divisive.  The offer is
an attack on the freedoms we stand for.  Why are we even debating it?

-- 
Alexandre Oliva, happy hacker            https://blog.lx.oliva.nom.br/
Free Software Activist     FSFLA co-founder     GNU Toolchain Engineer
Learn the truth about Richard Stallman at https://stallmansupport.org/

Re: Nominations for Anthropic’s Claude for Open Source program

[Eli Zaretskii via Gdb]

> From: Alexandre Oliva <oliva@gnu.org>
> Cc: carlos@redhat.com, mark@klomp.org, pinskia@gmail.com,
>         adhemerval.zanella@linaro.org, libc-alpha@sourceware.org,
>         gdb@sourceware.org
> Date: Thu, 18 Jun 2026 03:59:23 -0300
> 
> On Jun 18, 2026, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> >> Cc: Mark Wielaard <mark@klomp.org>, Andrew Pinski <pinskia@gmail.com>,
> >> Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>, glibc
> >> developers <libc-alpha@sourceware.org>, gdb developers <gdb@sourceware.org>
> >> Date: Wed, 17 Jun 2026 13:28:57 -0300
> >> From: Alexandre Oliva via Gdb <gdb@sourceware.org>
> >> 
> >> Now, if it was another mailing list member forwarding spam that
> >> advertised a 30-days-gratis bait for a service entirely misaligned with
> >> the purposes of GNU in general and of GNU libc specifically, to the
> >> point of opposing those purposes, would you still be favorable to it?
> 
> > That's a significant exaggeration
> 
> It's not.
> 
> The offer is promotional, for a short time, to attract new users.

That's plausibly Anthropic's motivation, but you accused David in
advertising, not Anthropic.  I don't care about what you say about
Anthropic, but I do care what you say about David and other members of
our community here.

> Indeed, it fits the pattern that companies use with influencers
> nowadays, offering them gratis access to a product for them to publish
> their use of the product/service and thereby promote the product/service
> to an even broader public.  Companies usually pay for that.

David is not a company, nor is he (AFAIK) in any way related to
Anthropic.

> Now, maybe you're not disputing that it was advertisement.
> 
> Maybe you're disputing that it's misaligned with GNU.

I'm disputing your uncalled-for attack on David, and similar attacks
by others in this discussion.

> > Not relevant.  Blocking spam and bots, regardless of the content of
> > those messages, is the job of the list moderators and they generally
> > do their job just fine so far.
> 
> They seem to have missed the post that started this thread

No, because it was a legitimate post that doesn't qualify as spam or
anything near it.  You may think otherwise, but you are very wrong.
Your demand to censor such messages is also in stark contrast the
democratic traditions of GNU.

> > if we cannot talk between ourselves in a kinder manner
> 
> Being kind encompasses not extending dangerous and harmful offers onto
> your colleagues.  Smoke your crack if you must, but please don't try to
> "help" others get hooked through "generous" offers in here.  It's not
> kind.  It's not cool.  It's detrimental.  It's divisive.  The offer is
> an attack on the freedoms we stand for.  Why are we even debating it?

Because you are viciously attacking one of us (and now also myself),
for no good reason, using a language that should not be heard here.
Please stop!  It is completely inappropriate and unbecoming!

Re: Nominations for Anthropic’s Claude for Open Source program

[Alexandre Oliva via Gdb]

On Jun 18, 2026, Eli Zaretskii <eliz@gnu.org> wrote:

> That's plausibly Anthropic's motivation, but you accused David in
> advertising, not Anthropic.

Yeah, just like the TV or influencer's channel (or radio, magazine,
newspaper, website, etc) who advertises someone else's product, what
they're doing is advertising.  Is this a language issue, maybe?

> I do care what you say about David and other members of
> our community here.

FWIW, David didn't even participate in this thread.  This one was
initiated by Carlos.  David started a twin thread elsewhere.  Yeah, that
confused me too.  Of course I raised the issue in David's thread too, as
soon as I noticed the duplication.

> David is not a company, nor is he (AFAIK) in any way related to
> Anthropic.

It's almost as if you were arguing that getting someone else to post
your unsolicited commercial email makes it no longer be spam.  I really
don't think this is the case.

> I'm disputing your uncalled-for attack on David

Erhm.  What attack?  There's no attack on David, or on Carlos, for that
matter.  What I'm doing is call out the initiation of the threads as
spam, and insisting that we shouldn't normalize or repeat that kind of
post, because it would be detrimental to our shared goals.

If I'm attacking anything, it's an idea, not a person.  But I really
don't feel that it's an attack, it's just a civil debate, even when I
use strong words.  It's clear that some people feel strongly about this
divisive issue, myself included, which might possibly twist some
perceptions.

> No, because it was a legitimate post that doesn't qualify as spam or
> anything near it.  You may think otherwise, but you are very wrong.

Oh, how kind of you to let me know the ultimate truth that emanates from
wherever you get it.  I'm not that privileged, I just happen to hold my
own opinions, and sometimes I try to express them to try to come to
agreement with others who don't seem to share them.  Because these are
shared media, the alternative to coming to an agreement is for either
opinion to be imposed on other participants, which, as you seem to
suggest, is not cool.


And I haven't even raised yet the issue of David and Carlos legitimacy
to pick who gets assigned slots.  Shouldn't that choice be for the
community to make democratically, as you say?  As the choice on whether
to pass the "generous offer"?  Or is this democracy you call for only
desirable when someone with access to the ultimate truth finds it
favorable?

-- 
Alexandre Oliva, happy hacker            https://blog.lx.oliva.nom.br/
Free Software Activist     FSFLA co-founder     GNU Toolchain Engineer
Learn the truth about Richard Stallman at https://stallmansupport.org/

Re: Nominations for Anthropic’s Claude for Open Source program

[Eli Zaretskii via Gdb]

> From: Alexandre Oliva <oliva@gnu.org>
> Cc: carlos@redhat.com, mark@klomp.org, pinskia@gmail.com,
>         adhemerval.zanella@linaro.org, libc-alpha@sourceware.org,
>         gdb@sourceware.org
> Date: Thu, 18 Jun 2026 10:43:24 -0300
> 
> On Jun 18, 2026, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> > That's plausibly Anthropic's motivation, but you accused David in
> > advertising, not Anthropic.
> 
> Yeah, just like the TV or influencer's channel (or radio, magazine,
> newspaper, website, etc) who advertises someone else's product, what
> they're doing is advertising.  Is this a language issue, maybe?

The original message was not a TV advertisement, either.  And no, it
isn't a language issue, not from my side anyway.

> > David is not a company, nor is he (AFAIK) in any way related to
> > Anthropic.
> 
> It's almost as if you were arguing that getting someone else to post
> your unsolicited commercial email makes it no longer be spam.

No commercial email was posted here in that message, none at all.

> > I'm disputing your uncalled-for attack on David
> 
> Erhm.  What attack?  There's no attack on David, or on Carlos, for that
> matter.  What I'm doing is call out the initiation of the threads as
> spam, and insisting that we shouldn't normalize or repeat that kind of
> post, because it would be detrimental to our shared goals.

You are attacking people by using denigrating language that
characterizes their actions.

> If I'm attacking anything, it's an idea, not a person.  But I really
> don't feel that it's an attack, it's just a civil debate, even when I
> use strong words.  It's clear that some people feel strongly about this
> divisive issue, myself included, which might possibly twist some
> perceptions.

I suggest that you re-read your own messages.  They tell a very
different story.  If you don't think so, maybe ask a friend to give
you an independent opinion.

> > No, because it was a legitimate post that doesn't qualify as spam or
> > anything near it.  You may think otherwise, but you are very wrong.
> 
> Oh, how kind of you to let me know the ultimate truth that emanates from
> wherever you get it.  I'm not that privileged, I just happen to hold my
> own opinions, and sometimes I try to express them to try to come to
> agreement with others who don't seem to share them.  Because these are
> shared media, the alternative to coming to an agreement is for either
> opinion to be imposed on other participants, which, as you seem to
> suggest, is not cool.

You are entitled to your opinions, but please make a point of
expressing them politely.

> And I haven't even raised yet the issue of David and Carlos legitimacy
> to pick who gets assigned slots.  Shouldn't that choice be for the
> community to make democratically, as you say?  As the choice on whether
> to pass the "generous offer"?  Or is this democracy you call for only
> desirable when someone with access to the ultimate truth finds it
> favorable?

We can discuss this, but it is a separate issue.

Re: Nominations for Anthropic’s Claude for Open Source program

[Jeffrey Walton via Gdb]

On Thu, Jun 18, 2026 at 9:44 AM Alexandre Oliva via Gdb
<gdb@sourceware.org> wrote:
>
> On Jun 18, 2026, Eli Zaretskii <eliz@gnu.org> wrote:
> [...]
>
> > I'm disputing your uncalled-for attack on David
>
> Erhm.  What attack?  There's no attack on David, or on Carlos, for that
> matter.  What I'm doing is call out the initiation of the threads as
> spam, and insisting that we shouldn't normalize or repeat that kind of
> post, because it would be detrimental to our shared goals.

Related (or maybe off-topic)...  I think this is one of the behavioral
attributes of Gen-Z and above.  The behavioral psychologists tell us
that kids raised on smartphones and social media can lack problem
resolution skills and they are hypersensitive to criticism.[1,2,3,4]
To them, nearly everyone who disagrees with them is an attack.  (Big
Tech acknowledges the problems, but disputes the cause).

[1] https://www.theatlantic.com/technology/archive/2024/03/teen-childhood-smartphone-use-mental-health-effects/677722/
[2] https://www.vox.com/24127431/smartphones-young-kids-children-parenting-social-media-teen-mental-health
[3] https://www.scientificamerican.com/article/social-media-algorithms-warp-how-people-learn-from-each-other/
[4] https://www.noemamag.com/social-media-messed-up-our-kids-now-it-is-making-us-ungovernable/

Jeff

Re: Nominations for Anthropic’s Claude for Open Source program

[Thomas Dineen via Gdb]

Polite words do not seem to work! So here it is:

This reflector is exclusively for the development of GDB Open source 
Software!!!

Take your AI garbage to the appropriate forum.



On 6/17/2026 10:27 PM, Eli Zaretskii via Gdb wrote:
>> Cc: Mark Wielaard <mark@klomp.org>, Andrew Pinski <pinskia@gmail.com>,
>>   Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>, glibc
>>   developers <libc-alpha@sourceware.org>, gdb developers <gdb@sourceware.org>
>> Date: Wed, 17 Jun 2026 13:28:57 -0300
>> From: Alexandre Oliva via Gdb <gdb@sourceware.org>
>>
>> Now, if it was another mailing list member forwarding spam that
>> advertised a 30-days-gratis bait for a service entirely misaligned with
>> the purposes of GNU in general and of GNU libc specifically, to the
>> point of opposing those purposes, would you still be favorable to it?
> That's a significant exaggeration, to put it mildly.  I invite you
> (and everyone else who wants to make up their minds independently) to
> re-read the original message posted by David.  I find no advertisement
> there, let alone spam.
>
> Here's that message again:
>
>    Anthropic is generously offering 6 months of Claude Max 20x to Free
>    and Open Source Software projects to support the FOSS
>    ecosystem. These tools can be incredibly helpful with reviewing
>    upstream code, evaluating the conformance of a particular interface
>    to a standard, or just for your own learning and exploring the
>    implementation of the project or subsystem.
>
>    If you are an established, active contributor to the GNU Toolchain
>    and are interested in having access to Claude Max 20x under
>    Anthropic’s Claude for Open Source program, please respond to this
>    email thread with a short description of how you would utilize the
>    access and the potential impact on the GNU Toolchain.  The program's
>    slots are limited and Carlos O'Donell and I plan to review the
>    proposals. Please keep in mind that these are individual grants from
>    Anthropic to you personally. If you are approved you’ll get an email
>    from Anthropic with further instructions.
>
>    While the current GNU Project policy is not to accept LLM-generated
>    content in the projects, that still leaves other uses like bug
>    triage, patch review, and research as ways to use these tools.
>
> This basically says that if someone wants to use the 6-month offer to
> work on GNU tools, they should describe their plans and apply for a
> slot in using the tools.  That's it.  It even explicitly says in a
> P.S. that "This is not an endorsement of Anthropic, Claude or LLMs".
>
> Please be fair and don't accuse people who post here in something they
> didn't say.  You don't have to agree, and you don't need to volunteer
> to use Anthropic's proposal, but let's have minimally-civil
> discussions here, okay?  Even your selection of
> intentionally-denigrating words ("spam", "losers", etc.) is highly
> inappropriate in civilized discussions, given what the original
> message actually said.
>
>> Will your stance change when bots subscribe to these lists, and then
>> start posting advertisement targeted at our developers but otherwise
>> bearing no relationship with the project?  Will you then call for such
>> spam to be blocked, or will we all have to endure it because some day
>> you decreed that it was ok?
> Not relevant.  Blocking spam and bots, regardless of the content of
> those messages, is the job of the list moderators and they generally
> do their job just fine so far.
>
>>> Where I draw the line, and where it can be drawn is in what we write down
>>> in our documentation, likewise in our process.
>> Our mailing lists are part of the project as well, and people who take
>> part in the project or even lead it are expected to behave in line with
>> the project, rather than against it, at the very least when
>> participating in project fora and activities.
>>
>> Advertising services antagonic to and destructive of user freedom is
>> really way out of line.
> Again, there was no advertising.  This is uncalled-for.  Please find
> more polite words to express your disagreement (if you must: after
> all, that was a call for volunteers, and you don't have to volunteer
> nor respond if you don't like the proposal).
>
> More generally, I'm disappointed by the hostility and the general
> style some people allow themselves to use in this discussion.  We
> don't deserve to call ourselves a community that promotes a better
> society if we cannot talk between ourselves in a kinder manner.  Shame
> on us!

Re: Nominations for Anthropic’s Claude for Open Source program

[Andy Wang via Gdb]

I'm surprised that I am reading the whole thread this beautiful Friday
afternoon.  It is a total waste of time. GNU is dead, its spirit is
gone and we should abandon it to welcome this AI world where we can do
anything with Agents.

On Thu, Jun 18, 2026 at 1:11 PM Thomas Dineen via Gdb
<gdb@sourceware.org> wrote:
>
> Polite words do not seem to work! So here it is:
>
> This reflector is exclusively for the development of GDB Open source
> Software!!!
>
> Take your AI garbage to the appropriate forum.
>
>
>
> On 6/17/2026 10:27 PM, Eli Zaretskii via Gdb wrote:
> >> Cc: Mark Wielaard <mark@klomp.org>, Andrew Pinski <pinskia@gmail.com>,
> >>   Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>, glibc
> >>   developers <libc-alpha@sourceware.org>, gdb developers <gdb@sourceware.org>
> >> Date: Wed, 17 Jun 2026 13:28:57 -0300
> >> From: Alexandre Oliva via Gdb <gdb@sourceware.org>
> >>
> >> Now, if it was another mailing list member forwarding spam that
> >> advertised a 30-days-gratis bait for a service entirely misaligned with
> >> the purposes of GNU in general and of GNU libc specifically, to the
> >> point of opposing those purposes, would you still be favorable to it?
> > That's a significant exaggeration, to put it mildly.  I invite you
> > (and everyone else who wants to make up their minds independently) to
> > re-read the original message posted by David.  I find no advertisement
> > there, let alone spam.
> >
> > Here's that message again:
> >
> >    Anthropic is generously offering 6 months of Claude Max 20x to Free
> >    and Open Source Software projects to support the FOSS
> >    ecosystem. These tools can be incredibly helpful with reviewing
> >    upstream code, evaluating the conformance of a particular interface
> >    to a standard, or just for your own learning and exploring the
> >    implementation of the project or subsystem.
> >
> >    If you are an established, active contributor to the GNU Toolchain
> >    and are interested in having access to Claude Max 20x under
> >    Anthropic’s Claude for Open Source program, please respond to this
> >    email thread with a short description of how you would utilize the
> >    access and the potential impact on the GNU Toolchain.  The program's
> >    slots are limited and Carlos O'Donell and I plan to review the
> >    proposals. Please keep in mind that these are individual grants from
> >    Anthropic to you personally. If you are approved you’ll get an email
> >    from Anthropic with further instructions.
> >
> >    While the current GNU Project policy is not to accept LLM-generated
> >    content in the projects, that still leaves other uses like bug
> >    triage, patch review, and research as ways to use these tools.
> >
> > This basically says that if someone wants to use the 6-month offer to
> > work on GNU tools, they should describe their plans and apply for a
> > slot in using the tools.  That's it.  It even explicitly says in a
> > P.S. that "This is not an endorsement of Anthropic, Claude or LLMs".
> >
> > Please be fair and don't accuse people who post here in something they
> > didn't say.  You don't have to agree, and you don't need to volunteer
> > to use Anthropic's proposal, but let's have minimally-civil
> > discussions here, okay?  Even your selection of
> > intentionally-denigrating words ("spam", "losers", etc.) is highly
> > inappropriate in civilized discussions, given what the original
> > message actually said.
> >
> >> Will your stance change when bots subscribe to these lists, and then
> >> start posting advertisement targeted at our developers but otherwise
> >> bearing no relationship with the project?  Will you then call for such
> >> spam to be blocked, or will we all have to endure it because some day
> >> you decreed that it was ok?
> > Not relevant.  Blocking spam and bots, regardless of the content of
> > those messages, is the job of the list moderators and they generally
> > do their job just fine so far.
> >
> >>> Where I draw the line, and where it can be drawn is in what we write down
> >>> in our documentation, likewise in our process.
> >> Our mailing lists are part of the project as well, and people who take
> >> part in the project or even lead it are expected to behave in line with
> >> the project, rather than against it, at the very least when
> >> participating in project fora and activities.
> >>
> >> Advertising services antagonic to and destructive of user freedom is
> >> really way out of line.
> > Again, there was no advertising.  This is uncalled-for.  Please find
> > more polite words to express your disagreement (if you must: after
> > all, that was a call for volunteers, and you don't have to volunteer
> > nor respond if you don't like the proposal).
> >
> > More generally, I'm disappointed by the hostility and the general
> > style some people allow themselves to use in this discussion.  We
> > don't deserve to call ourselves a community that promotes a better
> > society if we cannot talk between ourselves in a kinder manner.  Shame
> > on us!

Re: Nominations for Anthropic’s Claude for Open Source program

[Siddhesh Poyarekar]

On 16/06/2026 07:03, Mark Wielaard wrote:
> Just that the David and Carlos show seems to have this pattern of
> maximal drama. 
Please refrain from personal attacks.

Sid

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/15/26 5:11 PM, Carlos O'Donell wrote:
> On 6/15/26 7:39 AM, Mark Wielaard wrote:
> I will raise this with the WG lead Jonathan Wakely this Friday, and ask
> the WG's opinion if they believe I can be impartial.
This was raised as the first point of order in the GCC working group
for AI Policy chaired by Jonathan Wakely.

The Working Group reviewed the issue and decided I was able to
be impartial and continue my contribution to the policy.

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Siddhesh Poyarekar]

On 12/06/2026 19:11, Andrew Pinski wrote:
> All of this talk about llm and llm reviews is off-topic for this list. Including the original email is off-topic and should not have been sent.

How so?  Adhemerval's comment mentioned LLM use for patch reviews in 
other GNU projects, which is useful information for this developer 
community.

>    Can you step down from the llm gcc working group for this email since 
> it shows a conflict of interest of prompting one product.  Plus it shows 
> poor taste of sending out an email about a non free product to a mailing 
> list about glibc development.
Visible maintainers of FOSS projects routinely get these kinds of offers 
upfront and forwarding them to community members IMO doesn't mean 
anything more than simply making others aware.  I don't see a conflict 
of interest here, but that's for the WG to decide.

Is this marketing for Anthropic?  Yes, and it is the right of the 
members of this community to reject it outright, or use it to experiment 
if they wish to, without contributing AI generated legally significant 
code to the project.

FWIW, I have a strong personal preference against using frontier LLMs. 
Maybe I'm more acceptable as a LLM WG member because I align with your 
position? ;)

Sid

Re: Nominations for Anthropic’s Claude for Open Source program

[Collin Funk via Gdb]

Adhemerval Zanella Netto <adhemerval.zanella@linaro.org> writes:

> gnulib guys seem to be experimenting with LLM for patch review recently [1],
> and it has uncovered real regressions [2]. And I recall that their 
> experience with Coverity has been subpar [3], and I think they even started 
> disregarding it.
>
> Personally, I see value in adding an LLM to the patch review phase, especially 
> since we lack manpower and initial triage can help newcomers and uncover some 
> issues.

I don't have access to Gnulib's coverity, but GNU coreutils shows a lot
of Gnulib stuff anyways. My feeling is that it is useful, but one has to
do some work to filter out perfectly legitimate code (e.g., the coverity
machine complains about type limit comparisons which may be different on
less common platforms).

I don't disagree that LLM review is useful. However, I do want to state
my *personal* opinions here as I did in that Gnulib thread. Anthropic
(and OpenAI, which has a similar program) tend to play this program up
to show how altruistic and great they are. However, the 6 month limit
makes it seem like a strategic plan to gain future subscribers.
Obviously, I understand their employees have to eat. They should
absolutely ban users abusing generosity. However, it feels a bit shitty
given that their models would be relatively useless if they could not
train on and then regurgitate, likely violating licenses, free software
written by maintainers that they claim to be doing favors for.

Collin

Re: Nominations for Anthropic’s Claude for Open Source program

[Sam James via Gdb]

[-- Attachment #1: Type: text/plain, Size: 1738 bytes --]

Collin Funk <collin.funk1@gmail.com> writes:

> Adhemerval Zanella Netto <adhemerval.zanella@linaro.org> writes:
>
>> gnulib guys seem to be experimenting with LLM for patch review recently [1],
>> and it has uncovered real regressions [2]. And I recall that their 
>> experience with Coverity has been subpar [3], and I think they even started 
>> disregarding it.
>>
>> Personally, I see value in adding an LLM to the patch review phase, especially 
>> since we lack manpower and initial triage can help newcomers and uncover some 
>> issues.
>
> I don't have access to Gnulib's coverity, but GNU coreutils shows a lot
> of Gnulib stuff anyways. My feeling is that it is useful, but one has to
> do some work to filter out perfectly legitimate code (e.g., the coverity
> machine complains about type limit comparisons which may be different on
> less common platforms).
>

This is a common sentiment I hear with Coverity. Generally useful but
need people to do the work on reviewing findings and keep on top of that.

> I don't disagree that LLM review is useful. However, I do want to state
> my *personal* opinions here as I did in that Gnulib thread. Anthropic
> (and OpenAI, which has a similar program) tend to play this program up
> to show how altruistic and great they are. However, the 6 month limit
> makes it seem like a strategic plan to gain future subscribers.

It appears that it's also not some specific offer they've made to our
community: https://claude.com/contact-sales/claude-for-oss and
https://www.anthropic.com/claude-for-oss-terms.

I think it unfortunately acts as advertising. I am sure that was not the
intent but it seems to be the effect nonetheless :(

> [...]

sam

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 418 bytes --]

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/12/26 7:46 PM, Collin Funk wrote:
> Adhemerval Zanella Netto <adhemerval.zanella@linaro.org> writes:
> 
>> gnulib guys seem to be experimenting with LLM for patch review recently [1],
>> and it has uncovered real regressions [2]. And I recall that their
>> experience with Coverity has been subpar [3], and I think they even started
>> disregarding it.
>>
>> Personally, I see value in adding an LLM to the patch review phase, especially
>> since we lack manpower and initial triage can help newcomers and uncover some
>> issues.
> 
> I don't have access to Gnulib's coverity, but GNU coreutils shows a lot
> of Gnulib stuff anyways. My feeling is that it is useful, but one has to
> do some work to filter out perfectly legitimate code (e.g., the coverity
> machine complains about type limit comparisons which may be different on
> less common platforms).

Agreed. That's generally been the case with glibc too.

> I don't disagree that LLM review is useful. However, I do want to state
> my *personal* opinions here as I did in that Gnulib thread. Anthropic
> (and OpenAI, which has a similar program) tend to play this program up
> to show how altruistic and great they are. However, the 6 month limit
> makes it seem like a strategic plan to gain future subscribers.
> Obviously, I understand their employees have to eat. They should
> absolutely ban users abusing generosity. However, it feels a bit shitty
> given that their models would be relatively useless if they could not
> train on and then regurgitate, likely violating licenses, free software
> written by maintainers that they claim to be doing favors for.

I agree.

This feeling is real and it sucks.

If anything in the face of LLMs I want to double down on actively
mentoring more developers to FOSS development.

See my comments here:
https://fosstodon.org/@codonell/116743112568763059

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Andrew Pinski via Gdb]

On Mon, Jun 15, 2026 at 2:17 PM Carlos O'Donell <carlos@redhat.com> wrote:
> If anything in the face of LLMs I want to double down on actively
> mentoring more developers to FOSS development.

When a newcoming came in asking for things to do last month, what was
given is most likely overwheling for them and never responded again.
https://inbox.sourceware.org/libc-alpha/CAFPohwASPGQHreidwNTR3Eo81f_TfsajkF0sH2qZFKqSUQ47mg@mail.gmail.com/

I even pointed out in the news letter when referering to this email
that "this newsletter could use some easy issues of the week for glibc
and a mentor to reach out to for it":
https://inbox.sourceware.org/libc-alpha/CAHt=NyPLDnHzbsV8xc7TC9Us9Djg6sDMo+QqKpbg0Kgnp6MzRw@mail.gmail.com/

But nothing, not even saying let me get back to you on this. Not even
a word. So I don't see any double down on this at all from you. In
fact I just see inaction from you.

Thanks,
Andrea

Re: Nominations for Anthropic’s Claude for Open Source program

[Adhemerval Zanella Netto via Gdb]

On 15/06/26 18:35, Andrew Pinski wrote:
> On Mon, Jun 15, 2026 at 2:17 PM Carlos O'Donell <carlos@redhat.com> wrote:
>> If anything in the face of LLMs I want to double down on actively
>> mentoring more developers to FOSS development.
> 
> When a newcoming came in asking for things to do last month, what was
> given is most likely overwheling for them and never responded again.
> https://inbox.sourceware.org/libc-alpha/CAFPohwASPGQHreidwNTR3Eo81f_TfsajkF0sH2qZFKqSUQ47mg@mail.gmail.com/

Sorry if I was too brief and potentially not welcoming; I could have spent 
some time digging into potential work and writing down better ideas. So I 
will take the opportunity and write them down:

1. Testing 

The testing would aim to both improve libsupport (the support/ subfolder) by 
consolidating internal functions (as LTP does for its testing) and add new 
tests to cover missing symbols or functionality. We now have container support, 
so we can extend testing for functionality that requires a more complex system 
setup.

Also, check whether the tests can be improved in terms of coverage and/or time 
spent. I started focusing on this area, and it culminated in a patchset to improve 
test scalability [1]. One issue with this kind of work is that it requires a lot
of GNU Make knowledge, but we do have many internal examples, especially new 
tests, on how to improve coverage and write better tests. 

One possible subproject is to document how to properly write glibc tests using 
libsupport and/or adapt the remaining one to use it. There are still around 30 
tests that use the old test-skeleton.c - 12 in libio/, the rest scattered across 
stdlib, rt, resolv, posix, nptl, io. 


2. Benchmarking

Another possible project is to improve benchmarks (benchmark subfolder), which
is currently cumbersome, quite hard to read, and somewhat difficult to parse. 
It would be really interesting to follow something like Google Benchmark [2],
which has many terminal color gimmicks, but focuses on providing proper, 
meaningful information without the need for extra parsing or using the bench 
scripts.


3. Documentation

Document undocumented functions in the manual. The manual/ has 55 .texi files
and a long tail of installed functions with no entry. Pick a small family (e.g. 
some wcs*, mkostemp variants, newer C23 additions), write the texinfo. This 
also has zero ABI risk, and is easy to review.


4. Compiler-warning / new-GCC cleanups

Build with a newer GCC or stricter flags and fix the warnings in one subdirectory. 
Bounded, and a good way to learn the -Werror discipline glibc enforces. 

Another possibility is to check which backports are required to enable the old 
release branch to build cleanly with newer GCC versions.


5. Add missing header annotations

Check new gcc/clang annotations  (__wur, __nonnull, __attribute_malloc__) and
apply them where clearly correct. Small but needs care — must not introduce 
false-positive  warnings in the test suite, which would teach the newcomer to 
actually run make check.


6. Internal script cleanup

This might be a more contentious one, but I think we can reduce our build 
dependencies by at least rewriting the Perl script to either awk or Python.



There are small (1 to 2 weeks' work) projects that I think are good for newcomers. 
Below are some medium-sized (2-6 weeks' work) projects that require some more 
internal knowledge and research:


7. C23 conformance gaps 

The implementation of C23 is nearly complete, so the remaining work is mostly 
conformance testing and coverage: cross-check against Joseph Myers's C23 status 
tracker on libc-alpha, find the few stragglers, and write the missing conform/ 
tests. Good intro to the standards machinery without needing to design new ABI.


8. Resolver/getaddrinfo test coverage using support/resolv_test.h

Plenty of resolver behavior is under-tested; the framework lets a newcomer spin 
up a fake in-process DNS server. Medium because the domain is fiddly, but the 
tooling is already built.

And I can work on mentoring if someone is willing to work on any of this work.

> 
> I even pointed out in the news letter when referering to this email
> that "this newsletter could use some easy issues of the week for glibc
> and a mentor to reach out to for it":
> https://inbox.sourceware.org/libc-alpha/CAHt=NyPLDnHzbsV8xc7TC9Us9Djg6sDMo+QqKpbg0Kgnp6MzRw@mail.gmail.com/
> 

Well, I did reach out to you, but your reply made it seem like you would take 
care of it yourself [3]. In any case, if you are will I can help you with 
the glibc side of the weekly news.

> But nothing, not even saying let me get back to you on this. Not even
> a word. So I don't see any double down on this at all from you. In
> fact I just see inaction from you.
> 
> Thanks,
> Andrea

[1] https://patchwork.sourceware.org/project/glibc/list/?series=62266
[2] https://github.com/google/benchmark
[3] https://gcc.gnu.org/pipermail/gcc/2025-September/246616.html

Re: Nominations for Anthropic’s Claude for Open Source program

[Xi Ruoyao via Gdb]

On Mon, 2026-06-15 at 20:10 -0300, Adhemerval Zanella Netto wrote:
> 6. Internal script cleanup
> 
> This might be a more contentious one, but I think we can reduce our build 
> dependencies by at least rewriting the Perl script to either awk or Python.

Hmm, it seems glibc can already build w/o Perl (if you don't care the
info pages which need Texinfo).  I happened to figure that out after
forgetting the temporary Perl and Texinfo installations before building
glibc when I performed a test build of Linux From Scratch manually
several days ago.

Also reducing the Perl scripts won't help too much for distros as
Texinfo is still in Perl and the distros will need the info pages
anyway.

-- 
Xi Ruoyao <xry111@xry111.site>

Re: Nominations for Anthropic’s Claude for Open Source program

[Adhemerval Zanella Netto via Gdb]

On 16/06/26 11:50, Xi Ruoyao wrote:
> On Mon, 2026-06-15 at 20:10 -0300, Adhemerval Zanella Netto wrote:
>> 6. Internal script cleanup
>>
>> This might be a more contentious one, but I think we can reduce our build 
>> dependencies by at least rewriting the Perl script to either awk or Python.
> 
> Hmm, it seems glibc can already build w/o Perl (if you don't care the
> info pages which need Texinfo).  I happened to figure that out after
> forgetting the temporary Perl and Texinfo installations before building
> glibc when I performed a test build of Linux From Scratch manually
> several days ago.

We still install the mtrace, so to use all the glibc features one would
need to install perl. The scripts/test-installation.pl is already gated
through perl existence, but I think it is not a good design if we silent
disable testing due missing tools.

> 
> Also reducing the Perl scripts won't help too much for distros as
> Texinfo is still in Perl and the distros will need the info pages
> anyway.
> 

I think the main gain is for mtrace, the two other scripts
(test-installation.pl, summary.pl) are not essential.  The summary.pl is
only required for manual generation and you put texinfo already brings it
anyway.

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/16/26 10:59 AM, Adhemerval Zanella Netto wrote:
> 
> 
> On 16/06/26 11:50, Xi Ruoyao wrote:
>> On Mon, 2026-06-15 at 20:10 -0300, Adhemerval Zanella Netto wrote:
>>> 6. Internal script cleanup
>>>
>>> This might be a more contentious one, but I think we can reduce our build
>>> dependencies by at least rewriting the Perl script to either awk or Python.
>>
>> Hmm, it seems glibc can already build w/o Perl (if you don't care the
>> info pages which need Texinfo).  I happened to figure that out after
>> forgetting the temporary Perl and Texinfo installations before building
>> glibc when I performed a test build of Linux From Scratch manually
>> several days ago.
> 
> We still install the mtrace, so to use all the glibc features one would
> need to install perl. The scripts/test-installation.pl is already gated
> through perl existence, but I think it is not a good design if we silent
> disable testing due missing tools.

(1) Dependencies.

May we please start a new there for this?

Particularly around "build", "test", and "runtime" dependencies.
  
>>
>> Also reducing the Perl scripts won't help too much for distros as
>> Texinfo is still in Perl and the distros will need the info pages
>> anyway.
>>
> 
> I think the main gain is for mtrace, the two other scripts
> (test-installation.pl, summary.pl) are not essential.  The summary.pl is
> only required for manual generation and you put texinfo already brings it
> anyway.
> 

(2) Bootstrap.

May we please start a new thread for this?

This is another distinct discussion which has to do with reducing
toolchain bootstrap dependencies. I say "bootstrap" because you should
have a way to bootstrap (to do reproducible builds, or audit a minimal
bootstrap seed, or bringup hardware) that needs the fewest number of
things you can manage. This can include disabling features during the
bootstrap.

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Gokhan via Gdb]

It seems that we have both been trying to find the right way to move this
forward. I have been looking for ways to contribute, while you have been
trying to identify a suitable task for me.
I reviewed the areas you previously mentioned, but it seemed that I would
need to spend a significant amount of time getting familiar with them
first. For that reason, I postponed joining for a while.
This time, however, it looks like there may be some opportunities where I
can contribute more effectively. I will review the available options and
get back to you after I have had some time to assess them.
Please give me some time to look into this properly.

Kolaylıkla
*Kadir Gökhan Sezer*


On Tue, Jun 16, 2026 at 8:50 PM Carlos O'Donell <carlos@redhat.com> wrote:

> On 6/16/26 10:59 AM, Adhemerval Zanella Netto wrote:
> >
> >
> > On 16/06/26 11:50, Xi Ruoyao wrote:
> >> On Mon, 2026-06-15 at 20:10 -0300, Adhemerval Zanella Netto wrote:
> >>> 6. Internal script cleanup
> >>>
> >>> This might be a more contentious one, but I think we can reduce our
> build
> >>> dependencies by at least rewriting the Perl script to either awk or
> Python.
> >>
> >> Hmm, it seems glibc can already build w/o Perl (if you don't care the
> >> info pages which need Texinfo).  I happened to figure that out after
> >> forgetting the temporary Perl and Texinfo installations before building
> >> glibc when I performed a test build of Linux From Scratch manually
> >> several days ago.
> >
> > We still install the mtrace, so to use all the glibc features one would
> > need to install perl. The scripts/test-installation.pl is already gated
> > through perl existence, but I think it is not a good design if we silent
> > disable testing due missing tools.
>
> (1) Dependencies.
>
> May we please start a new there for this?
>
> Particularly around "build", "test", and "runtime" dependencies.
>
> >>
> >> Also reducing the Perl scripts won't help too much for distros as
> >> Texinfo is still in Perl and the distros will need the info pages
> >> anyway.
> >>
> >
> > I think the main gain is for mtrace, the two other scripts
> > (test-installation.pl, summary.pl) are not essential.  The summary.pl is
> > only required for manual generation and you put texinfo already brings it
> > anyway.
> >
>
> (2) Bootstrap.
>
> May we please start a new thread for this?
>
> This is another distinct discussion which has to do with reducing
> toolchain bootstrap dependencies. I say "bootstrap" because you should
> have a way to bootstrap (to do reproducible builds, or audit a minimal
> bootstrap seed, or bringup hardware) that needs the fewest number of
> things you can manage. This can include disabling features during the
> bootstrap.
>
> --
> Cheers,
> Carlos.
>
>

Re: Nominations for Anthropic’s Claude for Open Source program

[Joseph Myers via Gdb]

On Mon, 15 Jun 2026, Adhemerval Zanella Netto wrote:

> The testing would aim to both improve libsupport (the support/ subfolder) by 
> consolidating internal functions (as LTP does for its testing) and add new 
> tests to cover missing symbols or functionality. We now have container support, 
> so we can extend testing for functionality that requires a more complex system 
> setup.
> 
> Also, check whether the tests can be improved in terms of coverage and/or time 
> spent. I started focusing on this area, and it culminated in a patchset to improve 
> test scalability [1]. One issue with this kind of work is that it requires a lot
> of GNU Make knowledge, but we do have many internal examples, especially new 
> tests, on how to improve coverage and write better tests. 

I once generated lists of untested symbols in glibc by extracting 
non-compat symbols exported by glibc's shared libraries and looking for 
those that don't appear as dynamic symbol references in any binary in the 
glibc build tree after the testsuite was run.  (This is a very weak 
definition of being untested, in that a single reference to a symbol in a 
test binary suffices to make it count as tested.  Also, in principle 
compat symbols should be tested as well; those are just lower priority and 
more cumbersome to test.)

> 3. Documentation
> 
> Document undocumented functions in the manual. The manual/ has 55 .texi files
> and a long tail of installed functions with no entry. Pick a small family (e.g. 
> some wcs*, mkostemp variants, newer C23 additions), write the texinfo. This 
> also has zero ABI risk, and is easy to review.

scripts/documented.sh is *extremely* outdated (see the linuxthreads 
reference) but might be a basis for generating a newer list of 
undocumented functions.  Note that a lot of undocumented functions have 
bugs open in Bugzilla, so when documenting something, look to see if there 
is a bug to close.

> 7. C23 conformance gaps 
> 
> The implementation of C23 is nearly complete, so the remaining work is mostly 
> conformance testing and coverage: cross-check against Joseph Myers's C23 status 
> tracker on libc-alpha, find the few stragglers, and write the missing conform/ 
> tests. Good intro to the standards machinery without needing to design new ABI.

Collin should have C23 support for conform/, but maybe it was based on 
what glibc supported at a particular time rather than on exactly what's in 
the standard.

-- 
Joseph S. Myers
josmyers@redhat.com

Re: Nominations for Anthropic’s Claude for Open Source program

[Collin Funk via Gdb]

Joseph Myers <josmyers@redhat.com> writes:

>> 3. Documentation
>> 
>> Document undocumented functions in the manual. The manual/ has 55 .texi files
>> and a long tail of installed functions with no entry. Pick a small family (e.g. 
>> some wcs*, mkostemp variants, newer C23 additions), write the texinfo. This 
>> also has zero ABI risk, and is easy to review.
>
> scripts/documented.sh is *extremely* outdated (see the linuxthreads 
> reference) but might be a basis for generating a newer list of 
> undocumented functions.  Note that a lot of undocumented functions have 
> bugs open in Bugzilla, so when documenting something, look to see if there 
> is a bug to close.

+1, agreed.

>> 7. C23 conformance gaps 
>> 
>> The implementation of C23 is nearly complete, so the remaining work is mostly 
>> conformance testing and coverage: cross-check against Joseph Myers's C23 status 
>> tracker on libc-alpha, find the few stragglers, and write the missing conform/ 
>> tests. Good intro to the standards machinery without needing to design new ABI.
>
> Collin should have C23 support for conform/, but maybe it was based on 
> what glibc supported at a particular time rather than on exactly what's in 
> the standard.

I think I only made C23 equivalent to C11. Finishing the C23 conform
tests would be nice, but it is quite tedious. E.g., when I tried to add
stdint.h/inttypes.h conform tests, you mentioned that many platforms do
not yet support _BitInt [1]. There is another issue that not all
supported compilers support '-std=c23'. A new contributor could probably
figure all of this out, but I would be very surprised if they found it
enjoyable. :)

Collin

[1] https://inbox.sourceware.org/libc-alpha/fc312bc5-56b2-e9ee-6e89-476d02d91687@redhat.com/

Re: Nominations for Anthropic’s Claude for Open Source program

[Sam James via Gdb]

[-- Attachment #1: Type: text/plain, Size: 2099 bytes --]

Carlos O'Donell <carlos@redhat.com> writes:

> Developers,
>
> Anthropic is generously offering 6 months of Claude Max 20x to Free and Open
> Source Software projects to support the FOSS ecosystem. These tools can be
> incredibly helpful with reviewing upstream code, evaluating the conformance of
> a particular interface to a standard, or just for your own learning and
> exploring the implementation of the project or subsystem.

Several developers often don't even refer to Coverity, but say "a static
analyzer service" or similar. I think maybe some discussion should be
had before naming a service (and giving some publicity) given this
precedent, though it's a bit late now.

>
> If you are an established, active contributor to the GNU Toolchain and are
> interested in having access to Claude Max 20x under Anthropic’s Claude for Open
> Source program, please respond to this email thread with a short description of
> how you would utilize the access and the potential impact on the GNU Toolchain.
> The program's slots are limited and David Edelsohn and I plan to review the
> proposals. Please keep in mind that these are individual grants from Anthropic
> to you personally. If you are approved you’ll get an email from Anthropic with
> further instructions.
>
> While the current GNU Project policy is not to accept LLM-generated content in
> the projects, that still leaves other uses like bug triage, patch review, and
> research as ways to use these tools.

.. but also, I think we probably need to have a policy discussion about
this. I hope we won't have anyone pasting LLM replies onto bugs or patch
review, or an automated process doing that. I would *not* want to see
something like Sashiko for example. It would seriously put me off
contributing to glibc and I don't think it is respectful of
contributors' time who have come to us with a bug report or patch.

We should discuss this I think before handing out access.

Now, if a maintainer wants to use it to aid their review (with great
care), that may be more acceptable.

sam

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 418 bytes --]

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/12/26 7:12 PM, Sam James wrote:
> .. but also, I think we probably need to have a policy discussion about
> this. I hope we won't have anyone pasting LLM replies onto bugs or patch
> review, or an automated process doing that. I would *not* want to see
> something like Sashiko for example. It would seriously put me off
> contributing to glibc and I don't think it is respectful of
> contributors' time who have come to us with a bug report or patch.

100% agreed.

> We should discuss this I think before handing out access.

That's a good point.

> Now, if a maintainer wants to use it to aid their review (with great
> care), that may be more acceptable.
I absolutely agree that we need an AI policy for glibc ASAP.

I'm even happy to put in place stop-gap while the WG for GCC AI policy makes a first draft.

As a stop-gap I'd put up a "no LLM contributions accepted" policy in place.

We are also waiting on the GNU Project to finish policy review.

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Siddhesh Poyarekar]

On 15/06/2026 17:19, Carlos O'Donell wrote:
> I'm even happy to put in place stop-gap while the WG for GCC AI policy 
> makes a first draft.
> 
> As a stop-gap I'd put up a "no LLM contributions accepted" policy in place.
I'd be +1 for this, more specifically a policy similar to binutils. 
Others have expressed a desire to carve out an exception for testsuite, 
which is probably not that big a deal given that we have routinely 
incorporated tests from bugzilla without much regard for copyright 
ownership questions.

Sid

Re: Nominations for Anthropic’s Claude for Open Source program

[Joseph Myers via Gdb]

On Tue, 16 Jun 2026, Siddhesh Poyarekar wrote:

> On 15/06/2026 17:19, Carlos O'Donell wrote:
> > I'm even happy to put in place stop-gap while the WG for GCC AI policy makes
> > a first draft.
> > 
> > As a stop-gap I'd put up a "no LLM contributions accepted" policy in place.
> I'd be +1 for this, more specifically a policy similar to binutils. Others
> have expressed a desire to carve out an exception for testsuite, which is
> probably not that big a deal given that we have routinely incorporated tests
> from bugzilla without much regard for copyright ownership questions.

And for non-native speakers using LLMs to help translate their comments to 
English?

-- 
Joseph S. Myers
josmyers@redhat.com

Re: Nominations for Anthropic’s Claude for Open Source program

[Carlos O'Donell via Gdb]

On 6/16/26 12:36 PM, Joseph Myers wrote:
> On Tue, 16 Jun 2026, Siddhesh Poyarekar wrote:
> 
>> On 15/06/2026 17:19, Carlos O'Donell wrote:
>>> I'm even happy to put in place stop-gap while the WG for GCC AI policy makes
>>> a first draft.
>>>
>>> As a stop-gap I'd put up a "no LLM contributions accepted" policy in place.
>> I'd be +1 for this, more specifically a policy similar to binutils. Others
>> have expressed a desire to carve out an exception for testsuite, which is
>> probably not that big a deal given that we have routinely incorporated tests
>> from bugzilla without much regard for copyright ownership questions.
> 
> And for non-native speakers using LLMs to help translate their comments to
> English?
There is currently no exception for this activity, but I have added this to
the agenda to review for this Friday meeting of the WG for GCC AI policy.

-- 
Cheers,
Carlos.

Re: Nominations for Anthropic’s Claude for Open Source program

[Alexandre Oliva via Gdb]

On Jun 12, 2026, "Carlos O'Donell" <carlos@redhat.com> wrote:

> Anthropic is generously offering 6 months of Claude Max 20x to Free and Open
> Source Software projects to support the FOSS ecosystem.

You seem to have mispelled "poison" as "support".

I'd rather not contribute to this advertisement campaign for dis-SaaSS.
IMHO it doesn't belong in our communication channels.  At all.  It's an
embarrassment.

That said, since the harm is already done, I feel a duty to volunteer to
use up one of the available slots.

I have no intent whatsoever of using it.  I probably wouldn't be able to
as much as activate it anyway, as I'm sure it would involve some nonfree
software, in addition to SaaSS.

My purpose in applying for it is to spare someone else from being baited
into becoming a victim of this attack on our freedoms and on our ability
to think autonomously, and while at that to spare our tools from a
little additional pollution by these dis-services that would otherwise
likely take place.

I encourage others to apply with similar purposes, to defend our
freedom-loving communities from this attack aided by insiders.  Some
might wish to disguise their applications, to increase their odds of
going through.  This is self-defense against an unethical attack.

-- 
Alexandre Oliva, happy hacker            https://blog.lx.oliva.nom.br/
Free Software Activist     FSFLA co-founder     GNU Toolchain Engineer
Learn the truth about Richard Stallman at https://stallmansupport.org/