[maint] sim and common

[maint] sim and common

[Andrew Cagney]

Hello,

The sim/MAINTAINERS file contains:

>                 SIM Maintainers
> 
> The simulator is part of the GDB project, so see the file
> gdb/MAINTAINERS for general information about maintaining these files.
> 
> If you are considering contributing a patch, please see the file
> gdb/CONTRIBUTE.  Patches to these files should be posted to:
> gdb-patches@sources.redhat.com 
>         
>         Maintainers for particular sims:
> 
> arm             Nick Clifton <nickc@redhat.com>
> ppc             Andrew Cagney <ac131313@redhat.com>
> ppc             Geoff Keating <geoffk@redhat.com>
> mips            Andrew Cagney <ac131313@redhat.com>
> mips            Chris Demetriou <cgd@broadcom.com>
> common          Ben Elliston <bje@redhat.com>
> common          Frank Ch. Eigler <fche@redhat.com>
> 

(NickC worded this very well)

I'm wondering if it would be helpful for sim specific maintainers (at 
least for sim/common based sims) to have implicit approval/write 
permission on the sim/common directory.

Which reminds me, Stephane Carrez should really be listed as the m68hc11 
maintainer.  Unless, that is, GDB is going to assume that a GDB target 
maintainer implicitly maintains the corresponding SIM.

Thoughts?

Andrew

Re: [maint] sim and common

[cgd]

At Wed, 06 Mar 2002 00:08:27 -0500, Andrew Cagney wrote:
> I'm wondering if it would be helpful for sim specific maintainers (at 
> least for sim/common based sims) to have implicit approval/write 
> permission on the sim/common directory.

Uh, I was assuming that in general we (sim maintainers) had write
after approval perms, but that approval was still required.
(Concerned by your "write permission" comment there.  Do tell me if
i'm wrong.  8-)


Personally, I wouldn't mind someone else having responsibility for
reviewing my 'common' patches.

I'd like to make sure i'm not doing more damage than i intend, and i
say 'responsibility' because if somebody doesn't consider it their
responsibility well, then, it probably won't happen.

Of course, I'm sure we all have more than enough work to do...


My hope is to get approval, and hopefully consensus, for a bunch of
common patches over the next couple of months.  To "minor" things
like, oh, making the trace format a lot less lame.  8-)



chris

Re: [maint] sim and common

[Andrew Cagney]

> At Wed, 06 Mar 2002 00:08:27 -0500, Andrew Cagney wrote:
> 
>> I'm wondering if it would be helpful for sim specific maintainers (at 
>> least for sim/common based sims) to have implicit approval/write 
>> permission on the sim/common directory.
> 
> 
> Uh, I was assuming that in general we (sim maintainers) had write
> after approval perms, but that approval was still required.
> (Concerned by your "write permission" comment there.  Do tell me if
> i'm wrong.  8-)

Sim directory maintenance was, for a long time, very vague.  I'm trying 
to clarify things a little so that people that have a motivation to work 
in this directory can do so with little or no hinderance.  At the same 
time trying to ensure that a few basic procedures are followed.

I think part of this is ensuring that people directly affected by the 
infrastructure have a fairly free reign over it.

> Personally, I wouldn't mind someone else having responsibility for
> reviewing my 'common' patches.

Even if you have write privs on sim/common you're still free 
(encouraged?) to seek advice from others.  (Is this change a good idea? 
  In 20:20 hindsight was that bit a bad move?)

> I'd like to make sure i'm not doing more damage than i intend, and i
> say 'responsibility' because if somebody doesn't consider it their
> responsibility well, then, it probably won't happen.

Just as long as the damage is limited to ``flesh wounds'' :-)  Whats the 
worst that could happen - something gets broken?  Oops!  Actually, I 
think that the worst that could happen is for the code's development to 
stagnate (unless, that is, there is no one that thinks it is useful).

enjoy,
Andrew

Re: [maint] sim and common

[Andrew Cagney]

Did anyone else have a comment on this thread?

> At Wed, 06 Mar 2002 00:08:27 -0500, Andrew Cagney wrote:
> 
> I'm wondering if it would be helpful for sim specific maintainers (at least for sim/common based sims) to have implicit approval/write permission on the sim/common directory.
> 
> 
> Uh, I was assuming that in general we (sim maintainers) had write
> after approval perms, but that approval was still required.
> (Concerned by your "write permission" comment there.  Do tell me if
> i'm wrong.  8-)
> 
> Sim directory maintenance was, for a long time, very vague.  I'm trying to clarify things a little so that people that have a motivation to work in this directory can do so with little or no hinderance.  At the same time trying to ensure that a few basic procedures are followed.
> 
> I think part of this is ensuring that people directly affected by the infrastructure have a fairly free reign over it.
> 
> Personally, I wouldn't mind someone else having responsibility for
> reviewing my 'common' patches.
> 
> Even if you have write privs on sim/common you're still free (encouraged?) to seek advice from others.  (Is this change a good idea?  In 20:20 hindsight was that bit a bad move?)
> 
> I'd like to make sure i'm not doing more damage than i intend, and i
> say 'responsibility' because if somebody doesn't consider it their
> responsibility well, then, it probably won't happen.
> 
> Just as long as the damage is limited to ``flesh wounds'' :-)  Whats the worst that could happen - something gets broken?  Oops!  Actually, I think that the worst that could happen is for the code's development to stagnate (unless, that is, there is no one that thinks it is useful).
> 
> enjoy,
> Andrew

Re: [maint] sim and common

[Frank Ch. Eigler]

[-- Attachment #1: Type: text/plain, Size: 633 bytes --]

Hi -

On Sat, Mar 23, 2002 at 04:23:06PM -0500, Andrew Cagney wrote:
> Did anyone else have a comment on this thread?
> 
> > I'm wondering if it would be helpful for sim specific maintainers (at
> > least for sim/common based sims) to have implicit approval/write
> > permission on the sim/common directory.
> > [...]

Whether it is helpful or not can be ascertained when/if such maintainers
have patches for sim/common, and then if any "hindrance" in fact exists.
Preemptively granting such permission, in the absence of a history/intent
of specific contributions, goes against the grain of the rest of gdb.

- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [maint] sim and common

[Andrew Cagney]

> Hi -
> 
> On Sat, Mar 23, 2002 at 04:23:06PM -0500, Andrew Cagney wrote:
> 
>> Did anyone else have a comment on this thread?
>> 
> 
>> > I'm wondering if it would be helpful for sim specific maintainers (at
>> > least for sim/common based sims) to have implicit approval/write
>> > permission on the sim/common directory.
>> > [...]
> 
> 
> Whether it is helpful or not can be ascertained when/if such maintainers
> have patches for sim/common, and then if any "hindrance" in fact exists.
> Preemptively granting such permission, in the absence of a history/intent
> of specific contributions, goes against the grain of the rest of gdb.

If I understand the above correctly, your view is that, for the moment, 
only BenE and yourself should approve changes to src/sim/common?

enjoy,
Andrew

Re: [maint] sim and common

[Frank Ch. Eigler]

[-- Attachment #1: Type: text/plain, Size: 431 bytes --]

Hi -

On Mon, Mar 25, 2002 at 12:48:34PM -0500, Andrew Cagney wrote:
> If I understand the above correctly, your view is that, for the moment, 
> only BenE and yourself should approve changes to src/sim/common?

Not quite - the blanket write cabal can take action too.
Anyway, "for the moment" doesn't really matter: shelve
the issue until it is made topical by the appearance of
nontrivial sim/common patches.

- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [maint] sim and common

[cgd]

OK, well, what does this say for sim/igen?

(if the goal is to have a consistent policy of "patches are approved
by specific maintainers", then sim/igen needs some specific
maintainers.  8-)


chris

Re: [maint] sim and common

[Frank Ch. Eigler]

[-- Attachment #1: Type: text/plain, Size: 193 bytes --]

Hi -

On Mon, Mar 25, 2002 at 10:32:50AM -0800, cgd@broadcom.com wrote:
> OK, well, what does this say for sim/igen? [...]

IMO, you would be a good candidate to maintain sim/igen. :-)

- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [maint] sim and common

[cgd]

At Mon, 25 Mar 2002 13:49:18 -0500, Frank Ch. Eigler wrote:
> Hi -
> 
> On Mon, Mar 25, 2002 at 10:32:50AM -0800, cgd@broadcom.com wrote:
> > OK, well, what does this say for sim/igen? [...]
> 
> IMO, you would be a good candidate to maintain sim/igen. :-)

That's a nice thought, but I don't believe it's an accurate one.

I think that i've demonstrated well enough that I don't have enough
clue about igen to be allowed to maintain it.  8-)



chris

Re: [maint] sim and common

[Andrew Cagney]

> Hi -
> 
> On Mon, Mar 25, 2002 at 12:48:34PM -0500, Andrew Cagney wrote:
> 
>> If I understand the above correctly, your view is that, for the moment, 
>> only BenE and yourself should approve changes to src/sim/common?
> 
> 
> Not quite - the blanket write cabal can take action too.
> Anyway, "for the moment" doesn't really matter: shelve
> the issue until it is made topical by the appearance of
> nontrivial sim/common patches.

You mean?

     Blanket Write Privs
        (alphabetic)

Jim Blandy
Kevin Buettner
Andrew Cagney
J.T. Conklin
Fred Fish
Mark Kettenis
Peter Schauer
Stan Shebs
Michael Snyder
Elena Zannoni
Eli Zaretskii

Andrew

Re: [maint] sim and common

[Nick Clifton]

Hi Andrew,

> I'm wondering if it would be helpful for sim specific maintainers
> (at least for sim/common based sims) to have implicit approval/write
> permission on the sim/common directory.

I think that this would be a good idea - at least for target-specific
modifications to the common code base.

> Which reminds me, Stephane Carrez should really be listed as the 
> m68hc11 maintainer.  Unless, that is, GDB is going to assume that a 
> GDB target maintainer implicitly maintains the corresponding SIM.

I think that there probably is this assumption, but there is no harm
in making certain cases explicit, so I would say go ahead and do it.

Cheers
        Nick

Re: [maint] sim and common

[Ben Elliston]

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:

  Andrew> I'm wondering if it would be helpful for sim specific
  Andrew> maintainers (at least for sim/common based sims) to have
  Andrew> implicit approval/write permission on the sim/common
  Andrew> directory.

Why do you feel it would be helpful?  I don't think there has been any
evidence that patch approvals for sim/common has been a bottleneck or
indeed even a problem for anyone to date.

  Andrew> Which reminds me, Stephane Carrez should really be listed as
  Andrew> the m68hc11 maintainer.  Unless, that is, GDB is going to
  Andrew> assume that a GDB target maintainer implicitly maintains the
  Andrew> corresponding SIM.

I'm happy enough with that idea, provided that there is a mechanism
for people to be listed explicitly as sim maintainers, overriding the
corresponding GDB maintainer for that port.  There are potential sim
maintainers who are capable of working on the sim but might not be
willing/able to work on GDB as well.

Ben

Re: [maint] sim and common

[Andrew Cagney]

> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
> 
> 
> Andrew> I'm wondering if it would be helpful for sim specific
>   Andrew> maintainers (at least for sim/common based sims) to have
>   Andrew> implicit approval/write permission on the sim/common
>   Andrew> directory.
> 
> Why do you feel it would be helpful?  I don't think there has been any
> evidence that patch approvals for sim/common has been a bottleneck or
> indeed even a problem for anyone to date.

Is it unhelpful?  The people with the best idea for what to do with the 
common framework are most likely going to be those that are actively 
developing simulators.   Right now that is CGD (Chris D).

Anyway, further down in the thread, Frank has stated that, in his 
opinion, GDB's global write maintainers have ``global write'' on 
sim/common.  Is this what you understand?

(Unless otherwise stated, global-write stops when there is a maintainer).

>   Andrew> Which reminds me, Stephane Carrez should really be listed as
>   Andrew> the m68hc11 maintainer.  Unless, that is, GDB is going to
>   Andrew> assume that a GDB target maintainer implicitly maintains the
>   Andrew> corresponding SIM.
> 
> I'm happy enough with that idea, provided that there is a mechanism
> for people to be listed explicitly as sim maintainers, overriding the
> corresponding GDB maintainer for that port.

GDB's isa/abi, native, host and sim maintainers are recognized as 
separate independant roles (but are sometimes the same person).

The above was a suggestion for how to handle the situtation where the 
SIM role is vacent.  I think the consensus is that the SIM maintainers 
should be identified separatly and explicitly.

 > There are potential sim
> maintainers who are capable of working on the sim but might not be
> willing/able to work on GDB as well.

I don't understand.  Perhaphs you're thinking of a situtation like 
sim/arm/ or sim/ppc/.  There are two roles: GDB's sim maintainer (NickC, 
Cagney/GeoffK); and (independant of GDB) the original 
developers/contributors (Arm Ltd?, Cagney). (Yes, again, they can be the 
same person).

enjoy,
Andrew

Re: [maint] sim and common

[Ben Elliston]

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:

  >> Why do you feel it would be helpful?  I don't think there has been any
  >> evidence that patch approvals for sim/common has been a bottleneck or
  >> indeed even a problem for anyone to date.

  Andrew> Is it unhelpful?  The people with the best idea for what to do with the 
  Andrew> common framework are most likely going to be those that are actively 
  Andrew> developing simulators.   Right now that is CGD (Chris D).

While I agree with this point, I also like the idea of having a
devil's advocate to approve sim/common patches.  In the heat of the
moment, it's easy to think of patches to sim/common to solve
port-specific problems that are not in the best interest of all
simulators.

  Andrew> Anyway, further down in the thread, Frank has stated that,
  Andrew> in his opinion, GDB's global write maintainers have ``global
  Andrew> write'' on sim/common.  Is this what you understand?

I had not seen that, but okay.

  Andrew> The above was a suggestion for how to handle the situtation
  Andrew> where the SIM role is vacent.  I think the consensus is that
  Andrew> the SIM maintainers should be identified separatly and
  Andrew> explicitly.

Good.

Ben

Re: [maint] sim and common

[cgd]

At Tue, 9 Apr 2002 12:58:02 +0000 (UTC), "Ben Elliston" wrote:
> >>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
> 
>   >> Why do you feel it would be helpful?  I don't think there has been any
>   >> evidence that patch approvals for sim/common has been a bottleneck or
>   >> indeed even a problem for anyone to date.
> 
>   Andrew> Is it unhelpful?  The people with the best idea for what to do with the 
>   Andrew> common framework are most likely going to be those that are actively 
>   Andrew> developing simulators.   Right now that is CGD (Chris D).
> 
> While I agree with this point, I also like the idea of having a
> devil's advocate to approve sim/common patches.  In the heat of the
> moment, it's easy to think of patches to sim/common to solve
> port-specific problems that are not in the best interest of all
> simulators.

For what it's worth, I agree with Ben here.  8-)

I'd _prefer_ to have others review my changes (as long i can get that
review in a timely fashion)...


chris

Re: [maint] sim and common

[Ben Elliston]

>>>>> "cgd" == cgd  <cgd@broadcom.com> writes:

  >> While I agree with this point, I also like the idea of having a
  >> devil's advocate to approve sim/common patches.  In the heat of the
  >> moment, it's easy to think of patches to sim/common to solve
  >> port-specific problems that are not in the best interest of all
  >> simulators.

  cgd> For what it's worth, I agree with Ben here.  8-)

  cgd> I'd _prefer_ to have others review my changes (as long i can
  cgd> get that review in a timely fashion)...

Agreed--and as I stated, I don't believe there is any evidence to
suggest that timeliness is an issue at this point.

Ben

Re: [maint] sim and common

[Andrew Cagney]

> "cgd" == cgd  <cgd@broadcom.com> writes:
> 
> 
>>> While I agree with this point, I also like the idea of having a
>   >> devil's advocate to approve sim/common patches.  In the heat of the
>   >> moment, it's easy to think of patches to sim/common to solve
>   >> port-specific problems that are not in the best interest of all
>   >> simulators.
> 
>   cgd> For what it's worth, I agree with Ben here.  8-)
> 
>   cgd> I'd _prefer_ to have others review my changes (as long i can
>   cgd> get that review in a timely fashion)...
> 
> Agreed--and as I stated, I don't believe there is any evidence to
> suggest that timeliness is an issue at this point.

In addition to reviewing patches, a maintainer is typically assumed to 
be ``maintaining'' the existing code: fixing bugs, rewriting design 
mistakes, adding new features, ....

enjoy,
Andrew