Emacs Windows barebin distribution

Emacs Windows barebin distribution

[Christoph Scholtes]

Hi,

The purpose of the Emacs Windows barebin distribution has been 
questioned a couple of times. Does it still make sense to build it or 
can we save the time and server space?

Christoph

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> Date: Sun, 15 Apr 2012 08:28:12 -0600
> From: Christoph Scholtes <cschol2112@googlemail.com>
> 
> The purpose of the Emacs Windows barebin distribution has been 
> questioned a couple of times. Does it still make sense to build it or 
> can we save the time and server space?

I suggest to omit it the next time you produce binaries and see if
anyone complains.

Thanks.

Re: Emacs Windows barebin distribution

[Jambunathan K]

> The purpose of the Emacs Windows barebin distribution has been
> questioned a couple of times. Does it still make sense to build it or
> can we save the time and server space?

I will not complain if you drop it.

-- 

RE: Emacs Windows barebin distribution

[Drew Adams]

> I will not complain if you drop it.

Ditto.

Re: Emacs Windows barebin distribution

[Dani Moncayo]

[-- Attachment #1: Type: text/plain, Size: 622 bytes --]

On Sun, Apr 15, 2012 at 6:18 PM, Eli Zaretskii <eliz@gnu.org> wrote:
>> Date: Sun, 15 Apr 2012 08:28:12 -0600
>> From: Christoph Scholtes <cschol2112@googlemail.com>
>>
>> The purpose of the Emacs Windows barebin distribution has been
>> questioned a couple of times. Does it still make sense to build it or
>> can we save the time and server space?
>
> I suggest to omit it the next time you produce binaries and see if
> anyone complains.

AFAIK, nobody complained, and if nobody cares about that "barebin"
distribution, why spend time making it?

So please could we apply the attached patch?

Thanks.

-- 
Dani Moncayo

[-- Attachment #2: remove-barebin.diff --]
[-- Type: application/octet-stream, Size: 1117 bytes --]

=== modified file 'nt/zipdist.bat'
--- nt/zipdist.bat	2012-01-05 09:46:05 +0000
+++ nt/zipdist.bat	2012-11-16 22:47:25 +0000
@@ -25,7 +25,6 @@
 set TMP_DIST_DIR=emacs-%EMACS_VER%
 
 rem Check, if 7zip is installed and available on path
-:ZIP_CHECK
 7z
 if %ERRORLEVEL% NEQ 0 goto :ZIP_ERROR
 goto ZIP_DIST
@@ -35,14 +34,10 @@
 echo ERROR: Make sure 7zip is installed and available on the Windows Path!
 goto EXIT
 
-rem Build distributions
+rem Build and verify the binary distribution
 :ZIP_DIST
-rem Build and verify full distribution
 7z a -bd -tZIP -mx=9 -x!.bzrignore -x!.gitignore -xr!emacs.mdp -xr!*.pdb -xr!*.opt -xr!*~ -xr!CVS -xr!.arch-inventory emacs-%EMACS_VER%-bin-i386.zip %TMP_DIST_DIR%
 7z t emacs-%EMACS_VER%-bin-i386.zip
-rem Build and verify binary only distribution
-7z a -bd -tZIP -mx=9 -x!.bzrignore -x!.gitignore -xr!emacs.mdp -xr!*.pdb -xr!*.opt -xr!*~ -xr!CVS -xr!.arch-inventory emacs-%EMACS_VER%-barebin-i386.zip %TMP_DIST_DIR%/README.W32 %TMP_DIST_DIR%/bin %TMP_DIST_DIR%/etc/DOC-X %TMP_DIST_DIR%/COPYING
-7z t emacs-%EMACS_VER%-barebin-i386.zip
 goto EXIT
 
 :EXIT

Re: Emacs Windows barebin distribution

[Dani Moncayo]

[-- Attachment #1: Type: text/plain, Size: 215 bytes --]

> So please could we apply the attached patch?

I'm attaching a new version of the patch, which hides the output of 7z
when it is invoked only for the sake of checking if the program is
available.

-- 
Dani Moncayo

[-- Attachment #2: remove-barebin.diff --]
[-- Type: application/octet-stream, Size: 1131 bytes --]

=== modified file 'nt/zipdist.bat'
--- nt/zipdist.bat	2012-01-05 09:46:05 +0000
+++ nt/zipdist.bat	2012-11-16 22:51:10 +0000
@@ -25,8 +25,7 @@
 set TMP_DIST_DIR=emacs-%EMACS_VER%
 
 rem Check, if 7zip is installed and available on path
-:ZIP_CHECK
-7z
+7z 1>NUL
 if %ERRORLEVEL% NEQ 0 goto :ZIP_ERROR
 goto ZIP_DIST
 
@@ -35,14 +34,10 @@
 echo ERROR: Make sure 7zip is installed and available on the Windows Path!
 goto EXIT
 
-rem Build distributions
+rem Build and verify the binary distribution
 :ZIP_DIST
-rem Build and verify full distribution
 7z a -bd -tZIP -mx=9 -x!.bzrignore -x!.gitignore -xr!emacs.mdp -xr!*.pdb -xr!*.opt -xr!*~ -xr!CVS -xr!.arch-inventory emacs-%EMACS_VER%-bin-i386.zip %TMP_DIST_DIR%
 7z t emacs-%EMACS_VER%-bin-i386.zip
-rem Build and verify binary only distribution
-7z a -bd -tZIP -mx=9 -x!.bzrignore -x!.gitignore -xr!emacs.mdp -xr!*.pdb -xr!*.opt -xr!*~ -xr!CVS -xr!.arch-inventory emacs-%EMACS_VER%-barebin-i386.zip %TMP_DIST_DIR%/README.W32 %TMP_DIST_DIR%/bin %TMP_DIST_DIR%/etc/DOC-X %TMP_DIST_DIR%/COPYING
-7z t emacs-%EMACS_VER%-barebin-i386.zip
 goto EXIT
 
 :EXIT

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> Date: Fri, 16 Nov 2012 23:48:27 +0100
> From: Dani Moncayo <dmoncayo@gmail.com>
> Cc: Christoph Scholtes <cschol2112@googlemail.com>, emacs-devel@gnu.org
> 
> On Sun, Apr 15, 2012 at 6:18 PM, Eli Zaretskii <eliz@gnu.org> wrote:
> >> Date: Sun, 15 Apr 2012 08:28:12 -0600
> >> From: Christoph Scholtes <cschol2112@googlemail.com>
> >>
> >> The purpose of the Emacs Windows barebin distribution has been
> >> questioned a couple of times. Does it still make sense to build it or
> >> can we save the time and server space?
> >
> > I suggest to omit it the next time you produce binaries and see if
> > anyone complains.
> 
> AFAIK, nobody complained, and if nobody cares about that "barebin"
> distribution, why spend time making it?
> 
> So please could we apply the attached patch?

Fine with me.

Re: Emacs Windows barebin distribution

[Dani Moncayo]

[-- Attachment #1: Type: text/plain, Size: 289 bytes --]

>> So please could we apply the attached patch?
>
> Fine with me.

Thanks.

Since I can't commit to the trunk, could you do it?

(I'm attaching the patch again with yet another fix: when checking
whether 7z exists, hide the stderr stream in addition to stdout).

Thanks.

-- 
Dani Moncayo

[-- Attachment #2: remove-barebin.diff --]
[-- Type: application/octet-stream, Size: 1137 bytes --]

=== modified file 'nt/zipdist.bat'
--- nt/zipdist.bat	2012-01-05 09:46:05 +0000
+++ nt/zipdist.bat	2012-11-17 08:23:44 +0000
@@ -25,8 +25,7 @@
 set TMP_DIST_DIR=emacs-%EMACS_VER%
 
 rem Check, if 7zip is installed and available on path
-:ZIP_CHECK
-7z
+7z 1>NUL 2>NUL
 if %ERRORLEVEL% NEQ 0 goto :ZIP_ERROR
 goto ZIP_DIST
 
@@ -35,14 +34,10 @@
 echo ERROR: Make sure 7zip is installed and available on the Windows Path!
 goto EXIT
 
-rem Build distributions
+rem Build and verify the binary distribution
 :ZIP_DIST
-rem Build and verify full distribution
 7z a -bd -tZIP -mx=9 -x!.bzrignore -x!.gitignore -xr!emacs.mdp -xr!*.pdb -xr!*.opt -xr!*~ -xr!CVS -xr!.arch-inventory emacs-%EMACS_VER%-bin-i386.zip %TMP_DIST_DIR%
 7z t emacs-%EMACS_VER%-bin-i386.zip
-rem Build and verify binary only distribution
-7z a -bd -tZIP -mx=9 -x!.bzrignore -x!.gitignore -xr!emacs.mdp -xr!*.pdb -xr!*.opt -xr!*~ -xr!CVS -xr!.arch-inventory emacs-%EMACS_VER%-barebin-i386.zip %TMP_DIST_DIR%/README.W32 %TMP_DIST_DIR%/bin %TMP_DIST_DIR%/etc/DOC-X %TMP_DIST_DIR%/COPYING
-7z t emacs-%EMACS_VER%-barebin-i386.zip
 goto EXIT
 
 :EXIT

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> Date: Sat, 17 Nov 2012 09:27:47 +0100
> From: Dani Moncayo <dmoncayo@gmail.com>
> Cc: cschol2112@googlemail.com, emacs-devel@gnu.org
> 
> >> So please could we apply the attached patch?
> >
> > Fine with me.
> 
> Thanks.
> 
> Since I can't commit to the trunk, could you do it?

Done (trunk revision 110918).

Re: Emacs Windows barebin distribution

[Mathias Dahl]

[-- Attachment #1: Type: text/plain, Size: 572 bytes --]

Would it be possible to check the access logs to see how many people
actually download the barebin files?


On Sun, Apr 15, 2012 at 6:18 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> > Date: Sun, 15 Apr 2012 08:28:12 -0600
> > From: Christoph Scholtes <cschol2112@googlemail.com>
> >
> > The purpose of the Emacs Windows barebin distribution has been
> > questioned a couple of times. Does it still make sense to build it or
> > can we save the time and server space?
>
> I suggest to omit it the next time you produce binaries and see if
> anyone complains.
>
> Thanks.
>
>

[-- Attachment #2: Type: text/html, Size: 975 bytes --]

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> From: Mathias Dahl <mathias.dahl@gmail.com>
> Date: Sun, 18 Nov 2012 13:52:11 +0100
> Cc: Christoph Scholtes <cschol2112@googlemail.com>, emacs-devel@gnu.org
> 
> Would it be possible to check the access logs to see how many people
> actually download the barebin files?

Since I wrote the mail to which you responded, 7 months have passed
and no one complained.  I think by now it should be clear that no one
needs this.

RE: Emacs Windows barebin distribution

[Drew Adams]

> > Would it be possible to check the access logs to see how many people
> > actually download the barebin files?
> 
> Since I wrote the mail to which you responded, 7 months have passed
> and no one complained.  I think by now it should be clear that no one
> needs this.

No.  Bad reasoning.  Such silence does not at all make that clear.

1. I do not need/use barebin's, myself (to be clear).

2. Poll the _users_, as Richard so often requests in vain.  It means next to
nothing to wait 7 months and not hear any objection on the Emacs _development_
list.  This list does not represent a reasonable cross section of Emacs _users_.

3. Matthias's suggestion is the best approach, assuming it can be done.
Certainly the number of users who actually download barebins is an even more
accurate indication of the need than a user poll would be.

But my reason for chiming in here is to object to the point of view that a
7-month silence on the emacs-devel list is a good indication of what users want
or need.  That's, well, just silly.  And parochial.

Emacs is for its users, not for its developers.  That's the right attitude for
its developers to have, and it's the attitude that will give its developers the
most satisfaction in the long run.

Yes, Emacs has the advantage that many of its users also help develop it, and
its developers are also users.  But there are many other users out there, on
beyond emacs-devel.  And that includes users who download development versions
of the product.

Re: Emacs Windows barebin distribution

[Juanma Barranquero]

On Sun, Nov 18, 2012 at 4:50 PM, Drew Adams <drew.adams@oracle.com> wrote:

> Emacs is for its users, not for its developers.

On one hand, stopping releasing the barebin distribution does not
really harm (most?) users; there's still the full distribution.

On the other hand, it's not like removing the barebin target is hard
to revert, should the user's outcry suggest that it was indeed a bad
decision.

    Juanma

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> From: Juanma Barranquero <lekktu@gmail.com>
> Date: Sun, 18 Nov 2012 17:20:20 +0100
> Cc: Eli Zaretskii <eliz@gnu.org>, Mathias Dahl <mathias.dahl@gmail.com>, cschol2112@googlemail.com, 
> 	emacs-devel@gnu.org
> 
> On Sun, Nov 18, 2012 at 4:50 PM, Drew Adams <drew.adams@oracle.com> wrote:
> 
> > Emacs is for its users, not for its developers.
> 
> On one hand, stopping releasing the barebin distribution does not
> really harm (most?) users; there's still the full distribution.
> 
> On the other hand, it's not like removing the barebin target is hard
> to revert, should the user's outcry suggest that it was indeed a bad
> decision.

One important thing that the critics seem to forget, or don't know, is
that the barebin distribution simply didn't work, for quite some time
now.  So if we were to revert that decision, we would have to find a
way to make it work first.

RE: Emacs Windows barebin distribution

[Drew Adams]

> > Emacs is for its users, not for its developers.
>
> On one hand, stopping releasing the barebin distribution does not
> really harm (most?) users; there's still the full distribution.
> 
> On the other hand, it's not like removing the barebin target is hard
> to revert, should the user's outcry suggest that it was indeed a bad
> decision.

My point was about attitude and argument.  It was not about whether users would
actually be harmed by ending barebin distribution or whether it would be
difficult to undo any such harm.

Clearly, if the download data is available and easy to analyze, that would
pretty much answer the question of how much users really want/need this.  So I
supported Matthias's suggestion, if feasible.

But whether they do or do not does not change the falseness and blindness of the
rationale that just _because_ no users have chimed in on emacs-devel during the
last 7 months the user community must not want/need this.

Emacs-devel activity is not a very good measure of what users use/want/need.  It
is some measure, of course, but certainly not the best.

And in this case a very good measure was proposed.  And, unbelievably, it was
summarily dismissed/countered in favor of reliance on the much poorer measure of
emacs-devel activity.  That is the argument and the attitude I spoke out
against.

Re: Emacs Windows barebin distribution

[Juanma Barranquero]

On Sun, Nov 18, 2012 at 6:22 PM, Drew Adams <drew.adams@oracle.com> wrote:

> And in this case a very good measure was proposed.  And, unbelievably, it was
> summarily dismissed/countered in favor of reliance on the much poorer measure of
> emacs-devel activity.  That is the argument and the attitude I spoke out
> against.

Polling users doesn't happen on its own. Someone has to spend the time
to do it. Removing something deemed unnecessary and waiting for anyone
to complain is less work. That wouldn't be a good rationale for
killing a heavily used, or a core, feature, but it seems perfectly
reasonable to me when speaking of something that's likely not used by
anybody (more so after Eli's comment about it being broken).

So the argument and the attitude is, I think, "why should we spend
time in something that we're convinced it is useless anyway, unless we
have a good reason to think otherwise?"

    Juanma

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Cc: "'Eli Zaretskii'" <eliz@gnu.org>,
>         "'Mathias Dahl'" <mathias.dahl@gmail.com>, <cschol2112@googlemail.com>,
>         <emacs-devel@gnu.org>
> Date: Sun, 18 Nov 2012 09:22:01 -0800
> 
> My point was about attitude and argument.

There is no attitude.  The feature was broken for years, so no one
could have possibly used it.  Try giving us some credit before you
jump the gun.

RE: Emacs Windows barebin distribution

[Drew Adams]

> One important thing that the critics seem to forget, or don't know, is
> that the barebin distribution simply didn't work, for quite some time
> now.  So if we were to revert that decision, we would have to find a
> way to make it work first.

What critics?  Has anyone actually criticized or opposed discontinuing the
barebin distribution?  Straw man argument.

I spoke out against your argument for _how you know_ that users don't use
barebins.  I made it clear that I do not use barebins myself, and I do not in
any way oppose discontinuing their distribution.

It's the decision process that I spoke to: deciding what users want/need/use
based on emacs-devel list activity, instead of on a user poll or the proposed
analysis of download stats.

If you want to know what users use, ask them.
Or look directly at what they download.
If.  And I hope you do.

(That said, I do not take the position that whatever users might think they want
should determine the direction that Emacs Dev should go.  That would be tailism.
But there is a difference between (a) leading instead of following and (b)
ignoring those you are trying to lead.)

RE: Emacs Windows barebin distribution

[Drew Adams]

> > And in this case a very good measure was proposed.  And, 
> > unbelievably, it was summarily dismissed/countered in
> > favor of reliance on the much poorer measure of
> > emacs-devel activity.  That is the argument and the 
> > attitude I spoke out against.
> 
> Polling users doesn't happen on its own. Someone has to spend the time
> to do it.

The proposal from Matthias was not a proposal to poll the users.  He suggested
examining the download logs to see how much the barebins were picked up.

I do not know how much energy is required to do that, which is why my support
for it was qualified by "if easy", "if feasible" etc.

> Removing something deemed unnecessary and waiting for anyone
> to complain is less work. That wouldn't be a good rationale for
> killing a heavily used, or a core, feature, but it seems perfectly
> reasonable to me when speaking of something that's likely not used by
> anybody (more so after Eli's comment about it being broken).

I don't disagree with that at all.  I agree that it is a reasonable approach.

My argument was against the reasoning that _because_ no one has spoken up here
there must not be any user interest in this.  That's a false argument and
suggests a bad attitude, IMHO.

> So the argument and the attitude is, I think, "why should we spend
> time in something that we're convinced it is useless anyway, unless we
> have a good reason to think otherwise?"

See above.  That is a fair argument.  But that is not the argument and attitude
I was speaking about.

Re: Emacs Windows barebin distribution

[Juanma Barranquero]

On Sun, Nov 18, 2012 at 6:39 PM, Drew Adams <drew.adams@oracle.com> wrote:

> The proposal from Matthias was not a proposal to poll the users.  He suggested
> examining the download logs to see how much the barebins were picked up.

Fair enough. Still, that won't also happen on its own, and can be dead
easy, or not, depending on how accessible the logs are for a
sufficiently long interval of time.

> My argument was against the reasoning that _because_ no one has spoken up here
> there must not be any user interest in this.

Perhaps you're misinterpreting the reasoning, and it was more like "As
it has been broken for a long time, and we haven't received complains
on emacs-devel (or the bug list), it is reasonable to suppose that
there is no user interest in it."

> That's a false argument and suggests a bad attitude, IMHO.

With all due respect, I think you throw that accusation a bit lightly.

    Juanma

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Cc: <cschol2112@googlemail.com>, <emacs-devel@gnu.org>,
>         <mathias.dahl@gmail.com>
> Date: Sun, 18 Nov 2012 09:37:44 -0800
> 
> If you want to know what users use, ask them.

If a feature is broken and no one complains, I can know that without
asking.

RE: Emacs Windows barebin distribution

[Drew Adams]

> > The proposal from Matthias was not a proposal to poll the 
> > users.  He suggested examining the download logs to see
> > how much the barebins were picked up.
> 
> Fair enough. Still, that won't also happen on its own, and can be dead
> easy, or not, depending on how accessible the logs are for a
> sufficiently long interval of time.

Which is presumably why Mathias asked "would it be possible to check the access
logs?".  And it is why my support of the suggestion was qualified with "if
feasible", etc.  I have no idea how accessible or how accurate the logs are or
how easy it would be to check them.

> > My argument was against the reasoning that _because_ no one 
> > has spoken up here there must not be any user interest in this.
> 
> Perhaps you're misinterpreting the reasoning, and it was more like "As
> it has been broken for a long time, and we haven't received complains
> on emacs-devel (or the bug list), it is reasonable to suppose that
> there is no user interest in it."

To my knowledge, Eli's mention of it being broken came long after his argument
that I responded to.  His argument, and his dismissal of Mathias's suggestion
was just this - there is nothing here about anything being broken:

MD> > Would it be possible to check the access logs to see
MD> > how many people actually download the barebin files?
EZ> 
EZ> Since I wrote the mail to which you responded, 7 months
EZ> have passed and no one complained.  I think by now it
EZ> should be clear that no one needs this.

It is only that argument that I responded to.

> > That's a false argument and suggests a bad attitude, IMHO.
> 
> With all due respect, I think you throw that accusation a bit lightly.

I cannot know Eli's attitude, of course, which is why I said that his statement
_suggests_ such an attitude to me.  It is always a bad idea to attribute
motivations or feelings to others, I agree.

But to me the rationale he gave fits all too well with such an attitude.  That's
no proof, and anyway the point is not to prove anything about attitudes.

So let me just repeat that, regardless of what the attitudes might be here, we
should want to find out what users really use/want/need, and not just suppose
that emacs-devel activity reflects that.

Regardless of what one can guess wrt attitude, the argument/rationale itself
seems clear enough: lack of complaint here indicates that "no one needs this".

That was the argument given to reject Mathias's reasonable suggestion of a way
to see how many users "actually" use this.  And that argument was what I
objected to.

IOW, not only is it wrong (IMHO) to suppose that emacs-devel activity speaks
adequately for user needs.  It is also wrong to use that as a reason to
immediately reject a suggestion of a way to _really_ know those needs.

RE: Emacs Windows barebin distribution

[Drew Adams]

> > If you want to know what users use, ask them.
> 
> If a feature is broken and no one complains, I can know that without
> asking.

You are changing the subject.  My response to your argument rejecting checking
the logs was made long before you mentoned that the feature was broken.

At that point the discussion was about whether the barebins were being used.
You had not yet made clear (to my knowledge, at least) that there was anything
broken wrt barebins.

And your response to Mathias's suggestion to check the logs said nothing about
anything being broken.  You said only that no one had complained.

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Cc: <lekktu@gmail.com>, <cschol2112@googlemail.com>, <emacs-devel@gnu.org>,
>         <mathias.dahl@gmail.com>
> Date: Sun, 18 Nov 2012 10:22:44 -0800
> 
> > > If you want to know what users use, ask them.
> > 
> > If a feature is broken and no one complains, I can know that without
> > asking.
> 
> You are changing the subject.  My response to your argument rejecting checking
> the logs was made long before you mentoned that the feature was broken.

Which is exactly the point: time and again you allow yourself to pass
severe judgment on motives of others, without having _any_ idea
whatsoever about the facts of the matter, just because you smell some
"attitude".

If you want your opinions to matter, first learn the subject, or ask.
_Then_ state your opinion, based on facts.

> At that point the discussion was about whether the barebins were being used.
> You had not yet made clear (to my knowledge, at least) that there was anything
> broken wrt barebins.

I assume that people who write their opinions know what they are
talking about.  It is your duty to make yourself familiar with facts
before you speak up, especially when talking about something that was
last discussed 7 months ago.

Re: Emacs Windows barebin distribution

[Mathias Dahl]

[-- Attachment #1: Type: text/plain, Size: 1142 bytes --]

>
> > Would it be possible to check the access logs to see how many people
> > actually download the barebin files?
>
> Since I wrote the mail to which you responded, 7 months have passed
> and no one complained.  I think by now it should be clear that no one
> needs this.
>

I'm sorry for not noticing that. I don't know why Gmail decided to show me
that old e-mail all of a
sudden, I normally delete old e-mails from this list, not archive them, but
maybe "something"
happened.

I see there have been some arguing about your reply to my reply above,
and I'm not sure what to
say, so I'll try not to say too much about it ;) I just happened to see
your proposal of "testing" the
users and thought it would be much more natural to check the download logs,
which I assumed was
under control of whoever runs that server. A few "greps" (or whatever) and
some counts on those
logs and you would know, not for sure, but it would seem like a better
measurement than the planned
"test".

However, if the concept of the barebins were broken in some sense (that's
how I interpret your and
others' replies here), it's not anything worth arguing over.

[-- Attachment #2: Type: text/html, Size: 1624 bytes --]

Re: Emacs Windows barebin distribution

[Dani Moncayo]

For the record:

I'm in favor of removing the "barebin" distribution from the "dist"
makefile target, but actually I don't have anything against that
distribution.  What I wanted to achieve was just a way of building the
(full) "binary" distribution without having to build _also_ the
"barebin" one (which I'm not interested in, at least for now).

So, if for some reason it is deemed appropriate support the "barebin"
distribution again, please, make it a different makefile target, so
that those who don't want it don't have to spend the extra time and
disk space.

-- 
Dani Moncayo

Re: Emacs Windows barebin distribution

[Dani Moncayo]

> So, if for some reason it is deemed appropriate support the "barebin"
> distribution again, please, make it a different makefile target, so
> that those who don't want it don't have to spend the extra time and
> disk space.

Ah, BTW: Looking at how the barebin distribution was made, it seems
that it was simply the "bin" subdirectory of a normal (full) binary
distribution (well plus 3 documents).  If that is so, I see little
point in having a procedure (makefile target) for that; if someone is
interested, just download the normal distribution and pick up the
"bin" subdirectory.

-- 
Dani Moncayo

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> From: Mathias Dahl <mathias.dahl@gmail.com>
> Date: Sun, 18 Nov 2012 20:56:54 +0100
> Cc: Christoph Scholtes <cschol2112@googlemail.com>, emacs-devel@gnu.org
> 
> However, if the concept of the barebins were broken in some sense
> (that's how I interpret your and others' replies here), it's not
> anything worth arguing over.

It was broken in the sense that they couldn't be used for the purpose
for which they were invented, namely, to allow users re-dump Emacs
after loading some non-standard combination of Lisp packages.

The decision to drop the barebin distribution was made back in April,
we just wanted to wait for some time to see if someone will miss them,
after they were removed.  No one did.

Re: Emacs Windows barebin distribution

[Eli Zaretskii]

> Date: Sun, 18 Nov 2012 21:45:18 +0100
> From: Dani Moncayo <dmoncayo@gmail.com>
> Cc: Eli Zaretskii <eliz@gnu.org>, Christoph Scholtes <cschol2112@googlemail.com>, emacs-devel@gnu.org
> 
> Ah, BTW: Looking at how the barebin distribution was made, it seems
> that it was simply the "bin" subdirectory of a normal (full) binary
> distribution (well plus 3 documents).

You are looking at a broken barebin archive.  (Which doesn't surprise
me, since they've been broken for quite some time.)  There should be a
temacs.exe there and a dump.bat batch file, which allow one to re-dump
Emacs.  These are not in the normal binary distributions.

Re: Emacs Windows barebin distribution

[Stephen J. Turnbull]

Juanma Barranquero writes:
 > On Sun, Nov 18, 2012 at 6:22 PM, Drew Adams <drew.adams@oracle.com> wrote:
 > 
 > > And in this case a very good measure was proposed.  And,
 > > unbelievably, it was summarily dismissed/countered in favor of
 > > reliance on the much poorer measure of emacs-devel activity.
 > > That is the argument and the attitude I spoke out against.
 > 
 > Polling users doesn't happen on its own. Someone has to spend the time
 > to do it. Removing something deemed unnecessary and waiting for anyone
 > to complain is less work.

+1  Exactly what I would have said first.

I would have added that it's also far more accurate than polling,
though it (might) involve more pain for the users.

Re: Emacs Windows barebin distribution

[Jason Rumney]

"Drew Adams" <drew.adams@oracle.com> writes:

> 3. Matthias's suggestion is the best approach, assuming it can be done.
> Certainly the number of users who actually download barebins is an even more
> accurate indication of the need than a user poll would be.

The barebins were only ever useful for the following uses:

  1. Installing over the top of a source distribution without
     downloading the full bin package which has a lot of duplication
     with the source package.  In the days of 33k6 modems, this made
     some sense, but today the value is less obvious.

  2. Redumping Emacs. This use was questionable, as normally if Emacs
     were to be redumped, the user would need to change PURESIZE as
     well, so they would need to recompile everything themselves anyway.

I think today, these files are downloaded more in confusion over which
files are required, than by any deliberate need.  So studying the server
logs might not be as useful as you think.

Re: Emacs Windows barebin distribution

[Stephen J. Turnbull]

Jason Rumney writes:

 > I think today, these files are downloaded more in confusion over which
 > files are required, than by any deliberate need.  So studying the server
 > logs might not be as useful as you think.

Well, if you check for a following download of the whole shebang, say
within 24 hours, that would probably give you a pretty good indication
of just how large an attractive nuisance it was. :-)