[commit] Mention VAX floating-point support in NEWS

[commit] Mention VAX floating-point support in NEWS

[Mark Kettenis]

Thought this was worth mentioning.

Committed,

Mark


Index: ChangeLog
from  Mark Kettenis  <kettenis@gnu.org>

	* NEWS: Mention VAX floating point support.

Index: NEWS
===================================================================
RCS file: /cvs/src/src/gdb/NEWS,v
retrieving revision 1.175
diff -u -p -r1.175 NEWS
--- NEWS 4 Oct 2005 00:16:07 -0000 1.175
+++ NEWS 1 Nov 2005 07:27:10 -0000
@@ -56,6 +56,10 @@ After turning this setting "on", GDB pri
 when displaying arrays.  The default is "off" to preserve the previous
 behavior.
 
+* VAX floating point support
+
+GDB now supports the not-quite-ieee VAX F and D floating point formats.
+
 *** Changes in GDB 6.3:
 
 * New command line option

Re: [commit] Mention VAX floating-point support in NEWS

[Eli Zaretskii]

> Date: Tue, 1 Nov 2005 08:31:28 +0100 (CET)
> From: Mark Kettenis <mark.kettenis@xs4all.nl>
> 
> Thought this was worth mentioning.

It is, IMO.

> Committed,

Whatever happened to the custom to ask whether there are any
objections, or even (gasp!) make it an RFA?  NEWS is not one of the
files for which you are the responsible maintainer, and neither is
this change obvious, IMHO.

Re: [commit] Mention VAX floating-point support in NEWS

[Daniel Jacobowitz]

On Tue, Nov 01, 2005 at 09:54:17PM +0200, Eli Zaretskii wrote:
> > Date: Tue, 1 Nov 2005 08:31:28 +0100 (CET)
> > From: Mark Kettenis <mark.kettenis@xs4all.nl>
> > 
> > Thought this was worth mentioning.
> 
> It is, IMO.
> 
> > Committed,
> 
> Whatever happened to the custom to ask whether there are any
> objections, or even (gasp!) make it an RFA?  NEWS is not one of the
> files for which you are the responsible maintainer, and neither is
> this change obvious, IMHO.

From gdb/MAINTAINERS:

NEWS                    ALL

In any case, I don't see why additions to NEWS should require any more
approval than the feature they're describing.  Why isn't it obvious?

-- 
Daniel Jacobowitz
CodeSourcery, LLC

Re: [commit] Mention VAX floating-point support in NEWS

[Eli Zaretskii]

> Date: Tue, 1 Nov 2005 15:56:28 -0500
> From: Daniel Jacobowitz <drow@false.org>
> Cc: Mark Kettenis <mark.kettenis@xs4all.nl>, gdb-patches@sourceware.org
> 
> On Tue, Nov 01, 2005 at 09:54:17PM +0200, Eli Zaretskii wrote:
> > > Date: Tue, 1 Nov 2005 08:31:28 +0100 (CET)
> > > From: Mark Kettenis <mark.kettenis@xs4all.nl>
> > > 
> > > Thought this was worth mentioning.
> > 
> > It is, IMO.
> > 
> > > Committed,
> > 
> > Whatever happened to the custom to ask whether there are any
> > objections, or even (gasp!) make it an RFA?  NEWS is not one of the
> > files for which you are the responsible maintainer, and neither is
> > this change obvious, IMHO.
> 
> >From gdb/MAINTAINERS:
> 
> NEWS                    ALL

I don't know what this means, exactly (and MAINTAINERS doesn't
explain).  The practice until now was that changes to NEWS _were_
discussed.

Anyway, I think, after what's happened lately, we should try to err on
the side of more cooperation and collective decisions, not less.

> In any case, I don't see why additions to NEWS should require any more
> approval than the feature they're describing.  Why isn't it obvious?

It's not obvious because not every change we make goes into NEWS.

Re: [commit] Mention VAX floating-point support in NEWS

[Mark Kettenis]

> Date: Tue, 01 Nov 2005 21:54:17 +0200
> From: Eli Zaretskii <eliz@gnu.org>
>
> > Committed,
> 
> Whatever happened to the custom to ask whether there are any
> objections, or even (gasp!) make it an RFA?  NEWS is not one of the
> files for which you are the responsible maintainer, and neither is
> this change obvious, IMHO.

Sorry Eli. I didn't want to offend you.  But I thought I could just
add some blurb to NEWS for new features that went in.  Unless of
course there would be some controversy over whether it really is a new
feature or about the wording.  But I didn't think that was the case
for the VAX floating point support.

Mark

Re: [commit] Mention VAX floating-point support in NEWS

[Eli Zaretskii]

> Date: Tue, 1 Nov 2005 22:22:43 +0100 (CET)
> From: Mark Kettenis <mark.kettenis@xs4all.nl>
> CC: gdb-patches@sourceware.org
> 
> Sorry Eli. I didn't want to offend you.

I didn't take any offence.  Sorry if it sounded like I did.

> But I thought I could just add some blurb to NEWS for new features
> that went in.  Unless of course there would be some controversy over
> whether it really is a new feature or about the wording.  But I
> didn't think that was the case for the VAX floating point support.

I think we should carefully consider what goes into NEWS and how it is
worded.  NEWS is user-level documentation and should be treated as
such, IMO.

Re: [commit] Mention VAX floating-point support in NEWS

[Joel Brobecker]

> Whatever happened to the custom to ask whether there are any
> objections, or even (gasp!) make it an RFA?  NEWS is not one of the
> files for which you are the responsible maintainer, and neither is
> this change obvious, IMHO.

In my very humble opinion, I think it will overall cost less energy
if we let changes like this go in, and deal with the odd case where
the change is not obvious after all. My reasoning is that these cases
happen very seldomly as the global maintainers have demonstrated a good
sense of judgement, and it's always easy to revert if we need to discuss
this.

-- 
Joel

Re: [commit] Mention VAX floating-point support in NEWS

[Eli Zaretskii]

> Date: Tue, 1 Nov 2005 14:59:55 -0800
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: Mark Kettenis <mark.kettenis@xs4all.nl>,
> 	gdb-patches@sourceware.org
> 
> In my very humble opinion, I think it will overall cost less energy
> if we let changes like this go in, and deal with the odd case where
> the change is not obvious after all. My reasoning is that these cases
> happen very seldomly as the global maintainers have demonstrated a good
> sense of judgement, and it's always easy to revert if we need to discuss
> this.

I'm sorry to say, but your reasoning flies in the face of our recent
experience.  People (who will remain unnamed) were making
controversial changes to the repository, sometimes even over protests
of fellow maintainers.  As experience would have it, reverting is an
option that is seldom if ever taken; perhaps the reason is that
unilateral reverts will just start a revert war, because the person
whose patches are reverted has write access and can re-commit the
reverted patches.

Even if things don't go this far, reverting a change is something
deemed too radical and harsh (and justly so, IMHO), so I don't think
we should rely on it as part of our procedures.

Anyway, I'm amazed that people are so quick in forgetting the bitter
lessons we all should have learned from such recent events.

Re: [commit] Mention VAX floating-point support in NEWS

[Joel Brobecker]

> Anyway, I'm amazed that people are so quick in forgetting the bitter
> lessons we all should have learned from such recent events.

That's not correct. I didn't forget what happened in the past, but I
didn't come to the same conclusion as you did. I don't want to let the
example of one person (AFAIK) affect the productivity of the rest.
So far, I have been truly amazed at the positive attitude of all the
maintainers, and at the collaboration that we have.

My view is: Take a chance with your change. It's very easy to commit
a followup change if the wording needs to be improved or to revert it
if it turns out that it was improper. The thing is that it happens
infrequently, so overall, it's less work, so more time to improve GDB.

Now, that being said, I'm not a global maintainer, so it doesn't affect
me. If you guys prefer to enforce RFAs, I won't in the least bit object.
I was just giving my point of view on the matter.

Just as an aside, we at AdaCore use the review-after-commit approach.
This means that unless we're unsure of a change, we commit the change,
and the other engineers then review that change after. Very often the
change will be fine, but we have had occasions where we either had to
improve the change or even revert it. No hard feelings, no wars, it's
just part of human nature to sometimes fail.

-- 
Joel

Re: [commit] Mention VAX floating-point support in NEWS

[Eli Zaretskii]

> Date: Tue, 1 Nov 2005 21:18:02 -0800
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: mark.kettenis@xs4all.nl, gdb-patches@sourceware.org
> 
> I don't want to let the example of one person (AFAIK) affect the
> productivity of the rest.

I fail to see how asking for approval of a 2-liner, and a day or two
of delay waiting for responses, could possibly affect productivity.
For that matter, I don't see how changes in NEWS can affect
development in the first place.

> If you guys prefer to enforce RFAs, I won't in the least bit object.

I don't want to enforce RFAs, I'd like to see people do that out of
their free will.  Cooperation and comradeship cannot be enforced.

> Just as an aside, we at AdaCore use the review-after-commit approach.

There are many different modes of doing this.  Emacs, for example,
uses commit-at-will-without-any-review, and it works fairly well
(changes are almost never reverted).  But as long as we are following
the current GDB procedures and don't want to change them, we should
adhere to them, letter and spirit, because that's the only way of
restoring trust and good cooperative atmosphere.

Re: [commit] Mention VAX floating-point support in NEWS

[Jim Blandy]

Eli Zaretskii <eliz@gnu.org> writes:
> I fail to see how asking for approval of a 2-liner, and a day or two
> of delay waiting for responses, could possibly affect productivity.
> For that matter, I don't see how changes in NEWS can affect
> development in the first place.

It's three E-mails instead of one.  It's another pending change to
keep track of until the approval comes through.

But this wasn't really your point: if I'm reading your messages right,
you felt that 1) Joel was arguing that it was okay for Mark to
sidestep established practices because it would be more productive if
those practices where changed; and that 2) policies should be changed
through discussion and agreement, not just by following the policies
you wish you had.

I'll buy the second thesis, but I doubt the first.  Rather than seeing
Joel's mail as a post-facto justification for Mark's commit, why not
read it as a suggestion that we discuss and change the policy?
Looking at the archives, I think it reads just as well both ways.

<returning from the meta-meta-discussion to the meta-discussion>

I agree that NEWS is user-level documentation, and should be held to a
higher standard of general comprehensibility than, say, ChangeLog
entries.  I think there's a lot to be said for having a single person
in charge of editing the entries and keeping them grouped in some
logical fashion; when everyone just tacks stuff on at the front, it
gets kind of incoherent.  Given that Eli has been consistently
responsive to doc changes for years, I don't think this will be much
of a bottleneck.

<returning from the meta-discussion to the discussion>

Mark's change seems user-presentable and reasonably placed to me.

Re: [commit] Mention VAX floating-point support in NEWS

[Daniel Jacobowitz]

On Wed, Nov 02, 2005 at 02:42:24PM -0800, Jim Blandy wrote:
> I'll buy the second thesis, but I doubt the first.  Rather than seeing
> Joel's mail as a post-facto justification for Mark's commit, why not
> read it as a suggestion that we discuss and change the policy?
> Looking at the archives, I think it reads just as well both ways.

...

> I agree that NEWS is user-level documentation, and should be held to a
> higher standard of general comprehensibility than, say, ChangeLog
> entries.  I think there's a lot to be said for having a single person
> in charge of editing the entries and keeping them grouped in some
> logical fashion; when everyone just tacks stuff on at the front, it
> gets kind of incoherent.  Given that Eli has been consistently
> responsive to doc changes for years, I don't think this will be much
> of a bottleneck.

So how about this:

- For the moment, we change MAINTAINERS to mark NEWS as part of the
documentation.  Which seems plausible.

- In the longer term, we discuss the overall policy.

-- 
Daniel Jacobowitz
CodeSourcery, LLC

Re: [commit] Mention VAX floating-point support in NEWS

[Jim Blandy]

Daniel Jacobowitz <drow@false.org> writes:
> So how about this:
>
> - For the moment, we change MAINTAINERS to mark NEWS as part of the
> documentation.  Which seems plausible.
>
> - In the longer term, we discuss the overall policy.

Sounds good.

Re: [commit] Mention VAX floating-point support in NEWS

[Eli Zaretskii]

> Date: Wed, 2 Nov 2005 17:46:30 -0500
> From: Daniel Jacobowitz <drow@false.org>
> Cc: Eli Zaretskii <eliz@gnu.org>, Joel Brobecker <brobecker@adacore.com>,
> 	gdb-patches@sourceware.org
> 
> So how about this:
> 
> - For the moment, we change MAINTAINERS to mark NEWS as part of the
> documentation.  Which seems plausible.
> 
> - In the longer term, we discuss the overall policy.

Fine with me.