Re: The GNU Toolchain Infrastructure Project

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi,

We are using overseers to coordinate this and see how we can
mix-and-match pieces of this proposal. And to better understand how
this proposal interacts with Sourceware becoming a Conservancy member
project. So I added overseers@sourceware.org to have one central place
for these discussions.

On Wed, Sep 28, 2022 at 06:38:02PM -0400, Carlos O'Donell via Libc-alpha wrote:
> On 9/27/22 16:08, Carlos O'Donell wrote:
> > "The GNU Toolchain Infrastructure Project"
> > https://sourceware.org/pipermail/overseers/2022q3/018896.html
> 
> I've published the current GTI TAC meeting minutes to the glibc website:
> https://www.gnu.org/software/libc/gti-tac/index.html
> 
> The slides from the LF IT are a good overview:
> https://www.gnu.org/software/libc/gti-tac/LF%20IT%20Core%20Projects%20Services.pdf

I assume www.gnu.org was the easiest way for you to quickly make these
things public. But it now does look like it is an official FSF/GNU
proposal. Which I assume wasn't your intention. Note that it contains
a copyright notice "© 2022, GTI TAC." but doesn't seem to have a
(free) license. Which is kind of necessary if you host it on
www.gnu.org.

This does raise the question if you are also proposing migrating
non-sourceware services for projects like the websites of various of
the GNU projects on www.gnu.org or the release archives at the GNU ftp
server (and mirrors) those projects use.

The attendees list a subset of the GTI TAC members you posted
earlier. Was there any other way for people to participate in these
discussions? Did the GTI TAC invite the LF/IT team to give this
presentation or was this a proposal from the LF?

I note that this discussion and what you presented at Caudron was for
the migration of all services of all projects hosted on
Sourceware. But that your latest proposal is just for a subset of
projects, possibly only in part as would best suit their needs.

Lets file some sourceware infrastructure bugzilla entries for some of
these ideas in this presentation, to get a better understanding what
the real needs are. It would also be nice to hear the prices/budget
for the various options suggested in the presentation.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-02 16:47, Mark Wielaard via Overseers wrote:
>> I've published the current GTI TAC meeting minutes to the glibc website:
>> https://www.gnu.org/software/libc/gti-tac/index.html
>>
>> The slides from the LF IT are a good overview:
>> https://www.gnu.org/software/libc/gti-tac/LF%20IT%20Core%20Projects%20Services.pdf
> 
> I assume www.gnu.org was the easiest way for you to quickly make these
> things public. But it now does look like it is an official FSF/GNU
> proposal. Which I assume wasn't your intention. Note that it contains
> a copyright notice "© 2022, GTI TAC." but doesn't seem to have a
> (free) license. Which is kind of necessary if you host it on
> www.gnu.org.

Minutes moved here:

https://gti.gotplt.org/tac/
https://gti.gotplt.org/tac/LF%20IT%20Core%20Projects%20Services.pdf

I own gotplt.org and am happy to lend the subdomain for now to help 
coordinate this because I think the LF proposal is the best long term 
way forward for the GNU toolchain projects to remain competitive *and* Free.

To be clear, I don't think there are any qualms about adding a license 
notice here but we'd have to agree on one.  I made and shared this copy 
to dispel any further false speculation of scope creep of the GTI 
proposal.  For any content attributable to me in the meeting minutes, 
I'm happy to release it under any free license the TAC may agree on.

Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> [...] I think the LF proposal is the best long term way forward for
> the GNU toolchain projects to remain competitive *and* Free. [...]

Can you elaborate what risks in terms of competitiveness or freedom
you foresee with the status quo?  This is the first I recall hearing
of this concern.


- FChE

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-04 10:01, Frank Ch. Eigler wrote:
> Hi -
> 
>> [...] I think the LF proposal is the best long term way forward for
>> the GNU toolchain projects to remain competitive *and* Free. [...]
> 
> Can you elaborate what risks in terms of competitiveness or freedom
> you foresee with the status quo?  This is the first I recall hearing
> of this concern.

I don't see a risk to freedom.  The GNU toolchain is quite underfunded 
compared to llvm/clang and IMO it's a major risk to maintain status quo 
on that front.  The GTI opens new avenues for funding aspects of the GNU 
toolchain without affecting its core governance.

Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> > > [...] I think the LF proposal is the best long term way forward for
> > > the GNU toolchain projects to remain competitive *and* Free. [...]
> > 
> > Can you elaborate what risks in terms of competitiveness or freedom
> > you foresee with the status quo?  This is the first I recall hearing
> > of this concern.
> 
> I don't see a risk to freedom.  The GNU toolchain is quite underfunded
> compared to llvm/clang and IMO it's a major risk to maintain status quo on
> that front.  The GTI opens new avenues for funding aspects of the GNU
> toolchain without affecting its core governance.

What aspects of the gnu toolchain are open to being funded via the
LF/GTI proposal, -other than- the vast majority of the funds being
redirected to its own managed services infrastructure?

- FChE

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-04 10:19, Frank Ch. Eigler wrote:
>> I don't see a risk to freedom.  The GNU toolchain is quite underfunded
>> compared to llvm/clang and IMO it's a major risk to maintain status quo on
>> that front.  The GTI opens new avenues for funding aspects of the GNU
>> toolchain without affecting its core governance.
> 
> What aspects of the gnu toolchain are open to being funded via the
> LF/GTI proposal, -other than- the vast majority of the funds being
> redirected to its own managed services infrastructure?

This current proposal is limited to infrastructure, which has 
ever-growing needs.  Do you think the current proposal is not an upgrade 
to what we currently have?

Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> > > I don't see a risk to freedom.  The GNU toolchain is quite underfunded
> > > compared to llvm/clang and IMO it's a major risk to maintain status quo on
> > > that front.  The GTI opens new avenues for funding aspects of the GNU
> > > toolchain without affecting its core governance.
> > 
> > What aspects of the gnu toolchain are open to being funded via the
> > LF/GTI proposal, -other than- the vast majority of the funds being
> > redirected to its own managed services infrastructure?
> 
> This current proposal is limited to infrastructure, which has ever-growing
> needs.

I'm afraid I don't understand then what the point of comparing to LLVM
with respect to competitiveness or freedom was.  AIUI, infrastructure
is an enabler, not really a competitive differentiator.


> Do you think the current proposal is not an upgrade to what we
> currently have?

I don't know.  I am not under the impression that infrastructure is
holding back development on any of these projects.  Further, I suspect
that if the communities were given a choice to direct the sponsors'
generous donations toward new development type work, they may well
prefer that.  Is that possibility on offer?


- FChE

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-04 10:41, Frank Ch. Eigler wrote:
> I'm afraid I don't understand then what the point of comparing to LLVM
> with respect to competitiveness or freedom was.  AIUI, infrastructure
> is an enabler, not really a competitive differentiator.

I suppose that's a difference in our perception then.  I think of 
infrastructure as an accelerator and not just an enabler, which makes it 
a serious competitive differentiator.

>> Do you think the current proposal is not an upgrade to what we
>> currently have?
> 
> I don't know.  I am not under the impression that infrastructure is
> holding back development on any of these projects.  Further, I suspect
> that if the communities were given a choice to direct the sponsors'
> generous donations toward new development type work, they may well
> prefer that.  Is that possibility on offer?

Not in this proposal AFAICT (I have exactly the same information as you 
do) but IMO it would be great if it happens and the project communities 
accept it.

Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> > I'm afraid I don't understand then what the point of comparing to LLVM
> > with respect to competitiveness or freedom was.  AIUI, infrastructure
> > is an enabler, not really a competitive differentiator.
>
> I suppose that's a difference in our perception then.  I think of
> infrastructure as an accelerator and not just an enabler, which
> makes it a serious competitive differentiator.

Okay, we'd love to hear ideas for infrastructure changes that would
result in accelerating your work as developers.


- FChE

Re: The GNU Toolchain Infrastructure Project

[Alexandre Oliva]

On Oct  4, 2022, "Frank Ch. Eigler via Libc-alpha" <libc-alpha@sourceware.org> wrote:

> What aspects of the gnu toolchain are open to being funded via the
> LF/GTI proposal, -other than- the vast majority of the funds being
> redirected to its own managed services infrastructure?

Hear, hear,

I see a number of people, myself included, who are concerned that this
LF "offer" amounts to a power-grab, to use the "donations" as bait to
bring us into a trap in which our projects would be under control of a
body that has seats for sale, effectively "buying" the projects on the
cheap.


One way to significantly alleviate these concerns would be to test
whether the funds can be spent on infrastructure that's not under their
control, i.e., whether it's an investment, or possibly a gift that would
enable us to expand our autonomy rather than curtail it.

-- 
Alexandre Oliva, happy hacker                https://FSFLA.org/blogs/lxo/
   Free Software Activist                       GNU Toolchain Engineer
Disinformation flourishes because many people care deeply about injustice
but very few check the facts.  Ask me about <https://stallmansupport.org>

Re: The GNU Toolchain Infrastructure Project

[Christopher Faylor]

On Tue, Oct 04, 2022 at 09:46:08AM -0400, Siddhesh Poyarekar wrote:
>I made and shared this copy to dispel any further false speculation of
>scope creep of the GTI proposal.

Who is doing the false speculation?  Do you have a mailing list link?
It would be interesting to know who's got it wrong.

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-04 13:10, Christopher Faylor wrote:
> On Tue, Oct 04, 2022 at 09:46:08AM -0400, Siddhesh Poyarekar wrote:
>> I made and shared this copy to dispel any further false speculation of
>> scope creep of the GTI proposal.
> 
> Who is doing the false speculation?  Do you have a mailing list link?
> It would be interesting to know who's got it wrong.
> 

Mark asked upthread if content on gnu.org is also going to be migrated 
over based on sharing of meeting minutes on the gnu.org domain.

Sid

Re: The GNU Toolchain Infrastructure Project

[Christopher Faylor]

On Tue, Oct 04, 2022 at 01:17:14PM -0400, Siddhesh Poyarekar wrote:
>On 2022-10-04 13:10, Christopher Faylor wrote:
>> On Tue, Oct 04, 2022 at 09:46:08AM -0400, Siddhesh Poyarekar wrote:
>> > I made and shared this copy to dispel any further false speculation of
>> > scope creep of the GTI proposal.
>> 
>> Who is doing the false speculation?  Do you have a mailing list link?
>> It would be interesting to know who's got it wrong.
>
>Mark asked upthread if content on gnu.org is also going to be migrated over
>based on sharing of meeting minutes on the gnu.org domain.

I think you mean:

>>On Sun Oct 2 20:47:49 GMT 2022, Mark Wielaard wrote:
>This does raise the question if you are also proposing migrating
>non-sourceware services for projects like the websites of various of
>the GNU projects on www.gnu.org or the release archives at the GNU ftp
>server (and mirrors) those projects use.

Reading the meeting logs (I wasn't there and left this project shortly
after the meeting) I don't see anything that directly answers Mark's
question.  So, to me, this seems like an innocent request for
clarification rather than an attempt to push a false speculation.

There's no need to go down this rabbit hole, though.  Thanks for
clarifying.

cgf

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi Siddhesh,

On Tue, Oct 04, 2022 at 01:17:14PM -0400, Siddhesh Poyarekar wrote:
> On 2022-10-04 13:10, Christopher Faylor wrote:
> > On Tue, Oct 04, 2022 at 09:46:08AM -0400, Siddhesh Poyarekar wrote:
> > > I made and shared this copy to dispel any further false speculation of
> > > scope creep of the GTI proposal.
> > 
> > Who is doing the false speculation?  Do you have a mailing list link?
> > It would be interesting to know who's got it wrong.
> 
> Mark asked upthread if content on gnu.org is also going to be migrated over

I did indeed. Both the proposal and these minutes mention migrating
websites without mentioning any specifics. Knowing which websites are
meant and why they need migration is useful information.

The FSF tech team is helping us coordinating things on overseers to
help with release archives, mirroring, backups and sourceware
continuity. If this is about migrating websites currently on
www.gnu.org then talking to the FSF tech team does make sense. If it
isn't about that, then we will simply note that and move one.

We do take this proposal, and all other suggestions people make about
the sourceware infrastructure, seriously, but a lot of details of this
proposal are still unclear. We are trying to get as much details as
possible so we can see how this fits into the current sourceware
roadmap, get a better understanding of the budgetary needs, file
sourceware infrastructure bugs with those details. All to get a better
understanding what the real needs are and document the necessary steps
forward.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-04 15:05, Mark Wielaard wrote:
> I did indeed. Both the proposal and these minutes mention migrating
> websites without mentioning any specifics. Knowing which websites are
> meant and why they need migration is useful information.
> 
> The FSF tech team is helping us coordinating things on overseers to
> help with release archives, mirroring, backups and sourceware
> continuity. If this is about migrating websites currently on
> www.gnu.org then talking to the FSF tech team does make sense. If it
> isn't about that, then we will simply note that and move one.
> 
> We do take this proposal, and all other suggestions people make about
> the sourceware infrastructure, seriously, but a lot of details of this
> proposal are still unclear. We are trying to get as much details as
> possible so we can see how this fits into the current sourceware
> roadmap, get a better understanding of the budgetary needs, file
> sourceware infrastructure bugs with those details. All to get a better
> understanding what the real needs are and document the necessary steps
> forward.

I had in fact missed the websites mention, sorry I overreacted to your 
comment.  In that case, I don't know if the GNU websites are actually 
part of this proposal.

Sid

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi Siddhesh,

On Tue, Oct 04, 2022 at 03:10:35PM -0400, Siddhesh Poyarekar via Overseers wrote:
> > We do take this proposal, and all other suggestions people make about
> > the sourceware infrastructure, seriously, but a lot of details of this
> > proposal are still unclear. We are trying to get as much details as
> > possible so we can see how this fits into the current sourceware
> > roadmap, get a better understanding of the budgetary needs, file
> > sourceware infrastructure bugs with those details. All to get a better
> > understanding what the real needs are and document the necessary steps
> > forward.
> 
> I had in fact missed the websites mention, sorry I overreacted to your
> comment.  In that case, I don't know if the GNU websites are actually part
> of this proposal.

No worries. It seems everybody is somewhat unclear on the details of
this proposal. Making it hard for people not to speculate a
little. And it seems the scope changed between when various "key
stakeholders" were informed, the LF/IT presentation, the Cauldron talk
and what eventually got posted.

That is why we are trying to collect all details and file sourceware
infrastructure bugs to track the various technical needs from a
community perspective. But it would be really nice to hear directly
from the Linux Foundation and the OpenSSF about what exactly they are
proposing, which parts of the proposal are mandatory, which can be
mixed and matched, and how they see this working together with
Sourceware becoming a Software Freedom Conservancy member
project.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Christopher Faylor]

On Thu, Oct 06, 2022 at 10:02:19PM +0200, Mark Wielaard wrote:
>...But it would be really nice to hear directly from the Linux
>Foundation and the OpenSSF about what exactly they are proposing, which
>parts of the proposal are mandatory, which can be mixed and matched,
>and how they see this working together with Sourceware becoming a
>Software Freedom Conservancy member project.

Indeed.

The silence from the proponents of this project is puzzling.  I wonder
if this means there are more non-public negotiations going on somewhere,
leaving the community out of the loop.

cgf

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-06 16:12, Christopher Faylor via Overseers wrote:
> The silence from the proponents of this project is puzzling.  I wonder
> if this means there are more non-public negotiations going on somewhere,
> leaving the community out of the loop.

The proponents of this project are members of the GNU toolchain 
communities.  We approached the LF with the permission of the FSF to 
explore infrastructure funding solutions that would work for our 
communities.  The proposal has been made in response to our request, so 
we need to tell them what we need and not the other way around.

Also as I responded to Mark, the technical details of the transition are 
the responsibility of the GTI TAC (which you were invited to be member 
of and you declined) and not the LF IT, although they'd be the ones 
implementing and maintaining it.

We're at that stage at the moment where we look for consensus from the 
project communities so that we understand if we can move all of 
sourceware to LF IT or if we need both to coexist somehow.

Once we have a direction, we talk about what that transition would look 
like and ask questions accordingly.  Are there services that you 
absolutely cannot move to LF IT and why?  Why would you support (or 
oppose) porting the wiki to something like readthedocs backed by a git repo?

I respect your outright rejection of the proposal because at least it is 
clear that you don't have any stake in its fine tuning.

For everyone else, it's a proposal.  If there are changes you'd like to 
see in it, which will result in it being acceptable for you, please feel 
free to convey that.  If you think it is unnecessary for your project 
and that sourceware in its current state and vision is sufficient for 
your needs, please state that clearly too.

Sid

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi,

On Thu, 2022-10-06 at 17:37 -0400, Siddhesh Poyarekar wrote:
> Also as I responded to Mark, the technical details of the transition are 
> the responsibility of the GTI TAC (which you were invited to be member 
> of and you declined) and not the LF IT, although they'd be the ones 
> implementing and maintaining it.
> 
> We're at that stage at the moment where we look for consensus from the 
> project communities so that we understand if we can move all of 
> sourceware to LF IT or if we need both to coexist somehow.
> 
> Once we have a direction, we talk about what that transition would look 
> like and ask questions accordingly.  Are there services that you 
> absolutely cannot move to LF IT and why?  Why would you support (or 
> oppose) porting the wiki to something like readthedocs backed by a git repo?
> 
> I respect your outright rejection of the proposal because at least it is 
> clear that you don't have any stake in its fine tuning.

Lets try to make this a little less adversarial. This doesn't have to
be a clash of communities where there can be only one. Yes, the way
this was introduced caused things to become very contentious. But at
Cauldron we also agreed to bring this proposal to the overseers list
and discuss it together. Of course we can coexist. Lets do a reset. Now
that the plans are more public there will hopefully be less opportunity
for speculation and misunderstandings. But there are still some unclear
details and people have had various (unanswered) questions. It would be
good to get answers to the questions people asked on overseers. And it
would be great if the GTI TAC members discussed how they see the
technical details of various services on the overseers list. We can
then file specific sourceware infrastructure bugs to track the various
technical needs from a community perspective. And hopefully we can
then, as one community, take up shared responsibility of how to move
things forward together.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-06 16:02, Mark Wielaard wrote:
>> I had in fact missed the websites mention, sorry I overreacted to your
>> comment.  In that case, I don't know if the GNU websites are actually part
>> of this proposal.
> 
> No worries. It seems everybody is somewhat unclear on the details of
> this proposal. Making it hard for people not to speculate a
> little. And it seems the scope changed between when various "key
> stakeholders" were informed, the LF/IT presentation, the Cauldron talk
> and what eventually got posted.

I had not noticed the mention of websites in the proposal, which is why 
I was taken aback by its mention here.  That oversight is my fault, 
nothing to do with the proposal itself.

Could you clarify in what way you think the *scope* got changed between 
the private communications and the proposal that actually got posted? 
The technical details (which is different from scope) were never meant 
to be baked in, that's for the TAC to agree upon.  In that sense, the 
proposal details being open-ended is by design.

> That is why we are trying to collect all details and file sourceware
> infrastructure bugs to track the various technical needs from a

Fair enough.

> community perspective. But it would be really nice to hear directly
> from the Linux Foundation and the OpenSSF about what exactly they are
> proposing, which parts of the proposal are mandatory, which can be
> mixed and matched, and how they see this working together with
> Sourceware becoming a Software Freedom Conservancy member
> project.

You and others have been repeating "sourceware as a project" in a 
community owned sense as a truth for a while now but it really isn't. 
It is Red Hat owned infrastructure that is maintained by volunteers.  It 
is unquestioningly a community (and I'm proud part of it as someone who 
maintains the patchwork instance), but that's not the same thing as 
being an independent project that can do agreements and sign up for 
memberships.

Maybe Red Hat could spin off a sourceware project in some form that is 
actually capable of becoming a SFC member?  Or alternatively, 
"sourceware overseers" could become a body that maintains sourceware and 
is able to get funding through SFC for its activities?

Thanks,
Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> [...] Or alternatively, "sourceware overseers" could become a body
> that maintains sourceware and is able to get funding through SFC for
> its activities?

Great idea -- and this is roughly what's happening.  This "body"
consisting of key individuals has invited other folks interested in
helping with or helping guide the upkeep of shared sourceware
infrastructure to join us.

- FChE

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-06 17:36, Frank Ch. Eigler wrote:
>> [...] Or alternatively, "sourceware overseers" could become a body
>> that maintains sourceware and is able to get funding through SFC for
>> its activities?
> 
> Great idea -- and this is roughly what's happening.  This "body"
> consisting of key individuals has invited other folks interested in
> helping with or helping guide the upkeep of shared sourceware
> infrastructure to join us.

Here's another crazy idea on those lines then: how about having SFC fund 
sourceware overseers' time on TAC (in addition to, perhaps consulting 
tasks like independent security audits so that we have more eyes on the 
infrastructure) so that we continue to have them involved in the 
technical direction of GNU toolchain infrastructure?  That is of course 
for overseers who are actually able to accept payments from the SFC for 
such involvement.

Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> [...] so that we continue to have them involved in the technical
> direction of GNU toolchain infrastructure?  [...]

"continue"?  If the nature & degree of involvement we had so far in
the LF/GTI process is representative of the future, I'm not sure I can
in good faith ask anyone to fund our time on that.  Given that you are
listed as a TAC member, yet admitted being unclear on some details of
the proposal itself, perhaps we're in the same boat.

I cannot speak for the toolchain development community -- and have no
idea who honestly can -- but I suspect that some of the numerous
outstanding questions are material to their eventual decisionmaking on
moving their project to a new host - or staying.

- FChE

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-06 18:57, Frank Ch. Eigler wrote:
>> [...] so that we continue to have them involved in the technical
>> direction of GNU toolchain infrastructure?  [...]
> 
> "continue"?  If the nature & degree of involvement we had so far in
> the LF/GTI process is representative of the future, I'm not sure I can
> in good faith ask anyone to fund our time on that.  Given that you are
> listed as a TAC member, yet admitted being unclear on some details of
> the proposal itself, perhaps we're in the same boat.

Yes we are, in the sense that this is a proposal and the details are 
upon us (you're a listed member of TAC too) to help finalize.

Sid

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi Siddhesh,

On Thu, 2022-10-06 at 17:07 -0400, Siddhesh Poyarekar wrote:
> Could you clarify in what way you think the *scope* got changed
> between 
> the private communications and the proposal that actually got posted?

Given that they were private I can only talk for myself:
https://inbox.sourceware.org/overseers/Yz9dZWC9QIv+r4LH@elastic.org/T/#m22a52506bc116dbcb10c8cbfa8ed89510f4dc1b7
But various people listed as "key stakeholders consulted" said they
either didn't know anything about this, they were contacted but never
got any details, or were only told about parts of it.

> the proposal details being open-ended is by design.
> 
> > That is why we are trying to collect all details and file sourceware
> > infrastructure bugs to track the various technical needs from a
> 
> Fair enough.
> 
> > community perspective. But it would be really nice to hear directly
> > from the Linux Foundation and the OpenSSF about what exactly they are
> > proposing, which parts of the proposal are mandatory, which can be
> > mixed and matched, and how they see this working together with
> > Sourceware becoming a Software Freedom Conservancy member
> > project.
> 
> You and others have been repeating "sourceware as a project" in a 
> community owned sense as a truth for a while now but it really isn't. 
> It is Red Hat owned infrastructure that is maintained by volunteers.  It 
> is unquestioningly a community (and I'm proud part of it as someone who 
> maintains the patchwork instance), but that's not the same thing as 
> being an independent project that can do agreements and sign up for 
> memberships.
> [...]
> "sourceware overseers" could become a body that maintains sourceware and 
> is able to get funding through SFC for its activities?

That is precisely what we have been doing for the last couple of
months. And we don't believe that is in conflict with finding
alternative sources of funding, creating a technical advisory committee
and/or possibly having some "managed services" where that makes sense.

Some more background:

- Sourceware roadmap discussions
  https://sourceware.org/pipermail/overseers/2022q2/018453.html
  https://sourceware.org/pipermail/overseers/2022q2/018529.html
  https://sourceware.org/pipermail/overseers/2022q3/018636.html
  https://sourceware.org/pipermail/overseers/2022q3/018716.html
- Joining Software Freedom Conservancy as member project proposal
  https://sourceware.org/pipermail/overseers/2022q3/018802.html
- Full Sourceware SFC application text
  https://sourceware.org/pipermail/overseers/2022q3/018804.html
- Public SFC video chat meeting notes
  https://sourceware.org/pipermail/overseers/2022q3/018837.html
- Cauldron discussion notes and chat logs
  https://sourceware.org/pipermail/overseers/2022q3/018849.html

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-07 04:57, Mark Wielaard wrote:
> Given that they were private I can only talk for myself:
> https://inbox.sourceware.org/overseers/Yz9dZWC9QIv+r4LH@elastic.org/T/#m22a52506bc116dbcb10c8cbfa8ed89510f4dc1b7

I believe one of your concerns (alternatives for proprietary 
videoconferencing and list management for TAC) was acknowledged before 
you raised it in that email and is being worked on.

As for the rest, it really is a question on whether all of sourceware 
will in the end be migrated over to LF, it's for the remaining projects 
to decide.  If we indeed have all projects on board then I agree, 
perhaps we then need to ask Red Hat to hand sourceware over to the LF 
and call the project "Sourceware Infrastructure Project" or just Sourceware.

> But various people listed as "key stakeholders consulted" said they
> either didn't know anything about this, they were contacted but never
> got any details, or were only told about parts of it.

It would be nice to hear from these folks on what parts were withheld 
from them.

> That is precisely what we have been doing for the last couple of
> months. And we don't believe that is in conflict with finding
> alternative sources of funding, creating a technical advisory committee
> and/or possibly having some "managed services" where that makes sense.

That part is not in conflict, calling it the "Sourceware project" is.

Sid

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> [...]  As for the rest, it really is a question on whether all of
> sourceware will in the end be migrated over to LF, it's for the
> remaining projects to decide.  If we indeed have all projects on
> board [...]

"we" do not.  That option was taken off the table weeks ago.  For that
matter, I have not seen -any- project decisionmaking bodies formally
announce to/with their developer communities that they wished to move
away, only a few individuals who propose that this be done.

- FChE

Re: The GNU Toolchain Infrastructure Project

[Alexandre Oliva]

On Oct  7, 2022, Mark Wielaard <mark@klomp.org> wrote:

> Hi Siddhesh,
> On Thu, 2022-10-06 at 17:07 -0400, Siddhesh Poyarekar wrote:
>> Could you clarify in what way you think the *scope* got changed
>> between 
>> the private communications and the proposal that actually got posted?

> Given that they were private I can only talk for myself:
> https://inbox.sourceware.org/overseers/Yz9dZWC9QIv+r4LH@elastic.org/T/#m22a52506bc116dbcb10c8cbfa8ed89510f4dc1b7
> But various people listed as "key stakeholders consulted" said they
> either didn't know anything about this, they were contacted but never
> got any details, or were only told about parts of it.

That makes me very concerned.

Negotiating a community agreement in secrecy is worrysome to boot, but
giving different stakeholders different views of what the agreement
supposedly amounts to is a political trick normally used to push an
agreement through that would have been rejected by a majority, even if
for different reasons.  By presenting different views to different
parties, and misrepresenting their support for those partial views as
support for the whole they didn't even know about, one might put enough
pressure to persuade other parties to drop their objections, if they
believe the claimed broad support.

Even I got presented two very different views of the proposal by two of
its lead proponents, with different motivations (which is reasonable)
but factually conflicting commitments (which is not).

This all taken together makes me conclude that the alleged support for
the proposal, claimed by its lead proponents, is not something that can
be counted on, or taken for granted.  It needs to be double-checked by
circulating publicly a proposal encompassing everything that the
proposal entails, and then seeing whether it's actually acceptable as a
whole.  Given the chosen strategy, I suspect it won't be.

-- 
Alexandre Oliva, happy hacker                https://FSFLA.org/blogs/lxo/
   Free Software Activist                       GNU Toolchain Engineer
Disinformation flourishes because many people care deeply about injustice
but very few check the facts.  Ask me about <https://stallmansupport.org>

Re: The GNU Toolchain Infrastructure Project

[David Edelsohn]

[-- Attachment #1: Type: text/plain, Size: 3865 bytes --]

On Tue, Oct 11, 2022 at 12:00 PM Alexandre Oliva via Gcc <gcc@gcc.gnu.org>
wrote:

> On Oct  7, 2022, Mark Wielaard <mark@klomp.org> wrote:
>
> > Hi Siddhesh,
> > On Thu, 2022-10-06 at 17:07 -0400, Siddhesh Poyarekar wrote:
> >> Could you clarify in what way you think the *scope* got changed
> >> between
> >> the private communications and the proposal that actually got posted?
>
> > Given that they were private I can only talk for myself:
> >
> https://inbox.sourceware.org/overseers/Yz9dZWC9QIv+r4LH@elastic.org/T/#m22a52506bc116dbcb10c8cbfa8ed89510f4dc1b7
> > But various people listed as "key stakeholders consulted" said they
> > either didn't know anything about this, they were contacted but never
> > got any details, or were only told about parts of it.
>
> That makes me very concerned.
>
> Negotiating a community agreement in secrecy is worrysome to boot, but
> giving different stakeholders different views of what the agreement
> supposedly amounts to is a political trick normally used to push an
> agreement through that would have been rejected by a majority, even if
> for different reasons.  By presenting different views to different
> parties, and misrepresenting their support for those partial views as
> support for the whole they didn't even know about, one might put enough
> pressure to persuade other parties to drop their objections, if they
> believe the claimed broad support.
>

The "Sourceware as SFC member project proposal" has been negotiated in more
secrecy than the GTI proposal.  The "Sourceware" proposal was created
without input or support from key members of any of the GNU Toolchain
projects (GCC, GLIBC, GDB or Binutils).  The GTI proposal has been
circulated and socialized among the GNU Toolchain project leadership, GNU
Toolchain project Release Managers, key developers, active members of
"Overseers" and various stakeholders, including the FSF.

Where was a statement from key members of the GNU Toolchain projects -- the
people who actually use the services and infrastructure on a day to day
basis for their participation in the GNU Toolchain projects -- asking for
an alternative proposal? When were they allowed to participate in the
preparation of the "Sourceware" proposal, supposedly for their benefit?
All of the people with "skin in the game" who actively depend on the
services have been included and updated at each step of developing GTI, and
their feedback has helped shape the proposal.


>
> Even I got presented two very different views of the proposal by two of
> its lead proponents, with different motivations (which is reasonable)
> but factually conflicting commitments (which is not).
>

That is your assertion and accusation without any evidence.  Another
interpretation is that you didn't understand or you misinterpreted the
conversations.  Did you try to clarify this before making public
accusations?


>
> This all taken together makes me conclude that the alleged support for
> the proposal, claimed by its lead proponents, is not something that can
> be counted on, or taken for granted.  It needs to be double-checked by
> circulating publicly a proposal encompassing everything that the
> proposal entails, and then seeing whether it's actually acceptable as a
> whole.  Given the chosen strategy, I suspect it won't be.
>
> We appreciate everyone's opinion on this topic.

Those of us working on the GTI proposal have approached it with good
intentions and engaged everyone in good faith.  We have not made statements
maligning the motivations and intentions of those with different opinions,
implying nefarious motives, nor making baseless accusations.  We have been
open and available for conversations to clarify misunderstandings, and have
not used private conversations as public debating points nor for
divisive purposes.. I believe that speaks for itself.

Thanks, David

Re: The GNU Toolchain Infrastructure Project

[Frank Ch. Eigler]

Hi -

> [...] Where was a statement from key members of the GNU Toolchain
> projects -- the people who actually use the services and
> infrastructure on a day to day basis for their participation in the
> GNU Toolchain projects -- asking for an alternative proposal? When
> were they allowed to participate in the preparation of the
> "Sourceware" proposal, supposedly for their benefit?  [...]

This echoes a question asked during the Cauldron session.  I believe
it was during the second half, whose Zoom recording is for some reason
still not published.  Could you ask Jeremy to fix that please?

Anyway, to try to recount what I said then: the SFC proposal is
independent of the various guest projects.  It does not pretend to
speak for any of them.  It does not impose any changes on them.  All
the guests are just as welcome to come, stay, and leave, as they have
always been.  For this reason, it was not necessary to draw a
stakeholder map and conduct years-long negotiations behind the scenes.
Everyone has been invited to advise, in public, since August 30.


- FChE

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi David,

On Tue, Oct 11, 2022 at 01:14:50PM -0400, David Edelsohn wrote:
> an alternative proposal? When were they allowed to participate in the
> preparation of the "Sourceware" proposal, supposedly for their benefit?

It wasn't really meant as an alternative proposal. And tt shouldn't be
in conflict with finding alternative sources of funding, creating a
technical advisory committee or having some managed services. And it
is a about having a public discussion.

- Sourceware roadmap discussions
  https://sourceware.org/pipermail/overseers/2022q2/018453.html
  https://sourceware.org/pipermail/overseers/2022q2/018529.html
  https://sourceware.org/pipermail/overseers/2022q3/018636.html
  https://sourceware.org/pipermail/overseers/2022q3/018716.html
- Joining Software Freedom Conservancy as member project proposal
  https://sourceware.org/pipermail/overseers/2022q3/018802.html
- Full Sourceware SFC application text
  https://sourceware.org/pipermail/overseers/2022q3/018804.html
- Public SFC video chat meeting notes
  https://sourceware.org/pipermail/overseers/2022q3/018837.html
- Cauldron discussion notes and chat logs
  https://sourceware.org/pipermail/overseers/2022q3/018849.html

> Those of us working on the GTI proposal have approached it with good
> intentions and engaged everyone in good faith.  We have not made statements
> maligning the motivations and intentions of those with different opinions,
> implying nefarious motives, nor making baseless accusations.  We have been
> open and available for conversations to clarify misunderstandings

Then lets just let the past be the past. Now that the proposal is
public lets discuss it publicly. There have been various question
about the details on the overseers list. Lets just discuss those and
see how we can move forward.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Florian Weimer]

* Mark Wielaard via Overseers:

> Hi David,
>
> On Tue, Oct 11, 2022 at 01:14:50PM -0400, David Edelsohn wrote:
>> an alternative proposal? When were they allowed to participate in the
>> preparation of the "Sourceware" proposal, supposedly for their benefit?
>
> It wasn't really meant as an alternative proposal. And tt shouldn't be
> in conflict with finding alternative sources of funding, creating a
> technical advisory committee or having some managed services. And it
> is a about having a public discussion.
>
> - Sourceware roadmap discussions
>   https://sourceware.org/pipermail/overseers/2022q2/018453.html
>   https://sourceware.org/pipermail/overseers/2022q2/018529.html
>   https://sourceware.org/pipermail/overseers/2022q3/018636.html
>   https://sourceware.org/pipermail/overseers/2022q3/018716.html

Overseers was a hidden list until recently:

<https://web.archive.org/web/20220826033101/https://sourceware.org/mailman/listinfo>

I'm pointing this out to show how difficult it is to build public
consensus.  You might think you are doing it, but the view from the
outside is probably quite different.

Thanks,
Florian

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi Florian,

On Wed, Oct 12, 2022 at 03:18:55PM +0200, Florian Weimer wrote:
> * Mark Wielaard via Overseers:
> > And it is a about having a public discussion.
> >
> > - Sourceware roadmap discussions
> >   https://sourceware.org/pipermail/overseers/2022q2/018453.html
> >   https://sourceware.org/pipermail/overseers/2022q2/018529.html
> >   https://sourceware.org/pipermail/overseers/2022q3/018636.html
> >   https://sourceware.org/pipermail/overseers/2022q3/018716.html
> 
> Overseers was a hidden list until recently:
> 
> <https://web.archive.org/web/20220826033101/https://sourceware.org/mailman/listinfo>

Right, it wasn't advertised, but it was public:
https://web.archive.org/web/20220826033101/https://sourceware.org/mailman/listinfo/overseers

There were a couple of lists that were public, but not advertised,
which changed when we setup our public-inbox instance:
https://inbox.sourceware.org/overseers/YwJuV4e0I01sWxi0@wildebeest.org/

This was in part because we also handle account request on
overseers. It felt like a good idea to not make it easy for search
engines archive those. We now have a new (private, not archived)
account-requests list for that.

> I'm pointing this out to show how difficult it is to build public
> consensus.  You might think you are doing it, but the view from the
> outside is probably quite different.

Yes, I certainly see your point. But we did also post to the 20 most
active sourceware project lists about some proposals. And some of the
posts about the roadmap and the discussion about joining the
conservancy even made it to new sites like lwn:

Sourceware – GNU Toolchain Infrastructure roadmap
https://lwn.net/Articles/898655/
Sourceware seeking support from the Software Freedom Conservancy
https://lwn.net/Articles/906502/

And as the archives show we did publicly discuss things and actually
answered any questions people had:

- Joining Software Freedom Conservancy as member project proposal
  https://sourceware.org/pipermail/overseers/2022q3/018802.html
- Full Sourceware SFC application text
  https://sourceware.org/pipermail/overseers/2022q3/018804.html
- Public SFC video chat meeting notes
  https://sourceware.org/pipermail/overseers/2022q3/018837.html
- Cauldron discussion notes and chat logs
  https://sourceware.org/pipermail/overseers/2022q3/018849.html

I really liked some of these discussions. Hopefully in the future we
can do quarterly sourceware BBB video chats about any infrastructure
issues people/projects have.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Jonathan Corbet]

Mark Wielaard <mark@klomp.org> writes:

> Then lets just let the past be the past. Now that the proposal is
> public lets discuss it publicly. There have been various question
> about the details on the overseers list. Lets just discuss those and
> see how we can move forward.

Along those lines, I asked a few questions back in September:

  https://sourceware.org/pipermail/overseers/2022q3/018906.html

They seem relevant for anybody wanting to understand this proposal, and
the answers should be at the fingertips of the people putting it
together.  Any chance the rest of us could be enlightened?

Thanks,

jon

Re: The GNU Toolchain Infrastructure Project

[Alexandre Oliva]

On Oct 11, 2022, David Edelsohn <dje.gcc@gmail.com> wrote:

> open and available for conversations to clarify misunderstandings

Not useful when potential objectors are kept in the dark about the whole
thing.

> and have not used private conversations as public debating points nor for
> divisive purposes

The public claims of broad support used to put pressure for objectors to
give in seem to fit this pattern you deny, if not so much in seeding the
divide created by the then-secret proposal, but in bridging it.

The very purpose of private conversations was claimed by proponents of
the conversation as something to the effect of avoiding objections.

As for purporting key decisions as if in the hands of an advisory
committee, while the final decisions would rest in the hands of another
body whose members would be effectively buying the projects on the
cheap...

All of that, too, speaks for itself.


Anyway, this is all besides the point.  Whether or not there are
nefarious purposes behind it is besides the point.  The key point I
raise is that most people would support and accept something desirable
offered to them at no charge, but many might not upon finding that
there's a very steep price involved in the transaction.  There's no
evidence whatsoever that the costs have been conveyed along with the
dreams to the supposed supporters, so we'd better not take that alleged
support for granted.  The whole process was structured in a certain way,
explicitly for the purpose of sidelining objections.  That does not
inspire the very trust that would be required to agree to turn over
control over our infrastructure.

-- 
Alexandre Oliva, happy hacker                https://FSFLA.org/blogs/lxo/
   Free Software Activist                       GNU Toolchain Engineer
Disinformation flourishes because many people care deeply about injustice
but very few check the facts.  Ask me about <https://stallmansupport.org>

The GNU Toolchain Infrastructure Project

[Carlos O'Donell]

Community,

David Edelsohn and I are proud to announce and provide more detail about the 
GNU Toolchain Infrastructure project.

"The GNU Toolchain Infrastructure Project"
https://sourceware.org/pipermail/overseers/2022q3/018896.html

-- 
Cheers,
Carlos.

Re: The GNU Toolchain Infrastructure Project

[Carlos O'Donell]

On 9/27/22 16:08, Carlos O'Donell wrote:
> Community,
> 
> David Edelsohn and I are proud to announce and provide more detail about the 
> GNU Toolchain Infrastructure project.
> 
> "The GNU Toolchain Infrastructure Project"
> https://sourceware.org/pipermail/overseers/2022q3/018896.html
> 

I've published the current GTI TAC meeting minutes to the glibc website:
https://www.gnu.org/software/libc/gti-tac/index.html

The slides from the LF IT are a good overview:
https://www.gnu.org/software/libc/gti-tac/LF%20IT%20Core%20Projects%20Services.pdf

-- 
Cheers,
Carlos.

Re: The GNU Toolchain Infrastructure Project

[Nick Clifton]

Hi Everyone,

On 9/27/22 21:08, Carlos O'Donell via Binutils wrote:
> David Edelsohn and I are proud to announce and provide more detail about the
> GNU Toolchain Infrastructure project.

Just to be clear on my feelings, I believe that the position of the GNU Binutils
project should be that we are happy to support the GTI initiative, but we will
not be abandoning sourceware either.

There will not doubt be some technical issues to work through - arranging to
mirror the repositories for example - but I am sure that these can be resolved.

Cheers
   Nick

Re: The GNU Toolchain Infrastructure Project

[Mark Wielaard]

Hi Nick (CC overseers where we work out the technical details),

On Thu, Sep 29, 2022 at 11:02:44AM +0100, Nick Clifton via Binutils wrote:
> Just to be clear on my feelings, I believe that the position of the GNU Binutils
> project should be that we are happy to support the GTI initiative, but we will
> not be abandoning sourceware either.
> 
> There will not doubt be some technical issues to work through - arranging to
> mirror the repositories for example - but I am sure that these can be resolved.

Thanks for your feedback. I am sure we can arrange for binutils to
keep using the sourceware infrastructure while simultaneously work on
setting up mirrors of some of the services.

Cheers,

Mark

Re: The GNU Toolchain Infrastructure Project

[Carlos O'Donell]

On Thu, Sep 29, 2022 at 11:02:44AM +0100, Nick Clifton wrote:
> Hi Everyone,
> 
> On 9/27/22 21:08, Carlos O'Donell via Binutils wrote:
> > David Edelsohn and I are proud to announce and provide more detail about the
> > GNU Toolchain Infrastructure project.
> 
> Just to be clear on my feelings, I believe that the position of the GNU Binutils
> project should be that we are happy to support the GTI initiative, but we will
> not be abandoning sourceware either.

Thanks for the feedback Nick.

I encourage anyone else to respond to this thread and provide feedback.

It is certainly possible to have hybrid services, like you do already
with binutils (gnu.org website, and tarballs).

> There will not doubt be some technical issues to work through - arranging to
> mirror the repositories for example - but I am sure that these can be resolved.

Absolutely. Thanks for helping with the process.

Cheers,
Carlos.

Re: The GNU Toolchain Infrastructure Project

[Luis Machado]

On 9/29/22 11:02, Nick Clifton via Binutils wrote:
> Hi Everyone,
> 
> On 9/27/22 21:08, Carlos O'Donell via Binutils wrote:
>> David Edelsohn and I are proud to announce and provide more detail about the
>> GNU Toolchain Infrastructure project.
> 
> Just to be clear on my feelings, I believe that the position of the GNU Binutils
> project should be that we are happy to support the GTI initiative, but we will
> not be abandoning sourceware either.
> 
> There will not doubt be some technical issues to work through - arranging to
> mirror the repositories for example - but I am sure that these can be resolved.
> 
> Cheers
>    Nick
> 
> 

Binutils is tightly coupled to GDB though (or the other way around). I suppose both projects would
need to seek agreement on this, otherwise they'd have to be split at some point.

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-17 07:48, Luis Machado via Binutils wrote:
> On 9/29/22 11:02, Nick Clifton via Binutils wrote:
>> Hi Everyone,
>>
>> On 9/27/22 21:08, Carlos O'Donell via Binutils wrote:
>>> David Edelsohn and I are proud to announce and provide more detail 
>>> about the
>>> GNU Toolchain Infrastructure project.
>>
>> Just to be clear on my feelings, I believe that the position of the 
>> GNU Binutils
>> project should be that we are happy to support the GTI initiative, but 
>> we will
>> not be abandoning sourceware either.
>>
>> There will not doubt be some technical issues to work through - 
>> arranging to
>> mirror the repositories for example - but I am sure that these can be 
>> resolved.
>>
>> Cheers
>>    Nick
>>
>>
> 
> Binutils is tightly coupled to GDB though (or the other way around). I 
> suppose both projects would
> need to seek agreement on this, otherwise they'd have to be split at 
> some point.

I couldn't see any follow-up discussion on the gdb mailing list.  Have 
there been discussions elsewhere and is there a direction y'all have 
some consensus on?

Sid

Re: The GNU Toolchain Infrastructure Project

[Luis Machado]

On 10/17/22 13:08, Siddhesh Poyarekar wrote:
> On 2022-10-17 07:48, Luis Machado via Binutils wrote:
>> On 9/29/22 11:02, Nick Clifton via Binutils wrote:
>>> Hi Everyone,
>>>
>>> On 9/27/22 21:08, Carlos O'Donell via Binutils wrote:
>>>> David Edelsohn and I are proud to announce and provide more detail about the
>>>> GNU Toolchain Infrastructure project.
>>>
>>> Just to be clear on my feelings, I believe that the position of the GNU Binutils
>>> project should be that we are happy to support the GTI initiative, but we will
>>> not be abandoning sourceware either.
>>>
>>> There will not doubt be some technical issues to work through - arranging to
>>> mirror the repositories for example - but I am sure that these can be resolved.
>>>
>>> Cheers
>>>    Nick
>>>
>>>
>>
>> Binutils is tightly coupled to GDB though (or the other way around). I suppose both projects would
>> need to seek agreement on this, otherwise they'd have to be split at some point.
> 
> I couldn't see any follow-up discussion on the gdb mailing list.  Have there been discussions elsewhere and is there a direction y'all have some consensus on?
> 
> Sid

There have been none to my knowledge. And no consensus either.

Re: The GNU Toolchain Infrastructure Project

[Siddhesh Poyarekar]

On 2022-10-17 08:16, Luis Machado wrote:
>>> Binutils is tightly coupled to GDB though (or the other way around). 
>>> I suppose both projects would
>>> need to seek agreement on this, otherwise they'd have to be split at 
>>> some point.
>>
>> I couldn't see any follow-up discussion on the gdb mailing list.  Have 
>> there been discussions elsewhere and is there a direction y'all have 
>> some consensus on?
>>
>> Sid
> 
> There have been none to my knowledge. And no consensus either.
> 

The FSF is hosting a call for this next week, so hopefully that'll help 
the gdb community start a discussion and come to some consensus:

https://sourceware.org/pipermail/overseers/2022q4/018997.html

Sid