Windows emacs-25.1 i686 vs x86_64?

Windows emacs-25.1 i686 vs x86_64?

[David M. Miller]

When I went to download 25.1 for windows I saw two packages on the main GNU ftp server:

    emacs-25.1-i686-w64-mingw32.zip
     - AND -
    emacs-25.1-x86_64-w64-mingw32.zip

but no explanation of the difference between them in the README file.

Also, there seems to be a separate package of "deps" for each:

    emacs-25-i686-deps.zip
     - AND -
    emacs-25-x86_64-deps.zip

the need for which is also not explained anywhere (dependencies? dependencies for what?).

Does anyone know what the difference between the i686 and x86_64 packages are?
Does anyone know what the "deps" packages are for?

Thanks,
David

Re: Windows emacs-25.1 i686 vs x86_64?

[tomas]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, Nov 03, 2016 at 08:32:40AM -0700, David M. Miller wrote:
> When I went to download 25.1 for windows I saw two packages on the main GNU ftp server:
> 
>     emacs-25.1-i686-w64-mingw32.zip
>      - AND -
>     emacs-25.1-x86_64-w64-mingw32.zip
> 
> but no explanation of the difference between them in the README file.
> 
> Also, there seems to be a separate package of "deps" for each:
> 
>     emacs-25-i686-deps.zip
>      - AND -
>     emacs-25-x86_64-deps.zip
> 
> the need for which is also not explained anywhere (dependencies? dependencies for what?).
> 
> Does anyone know what the difference between the i686 and x86_64 packages are?
> Does anyone know what the "deps" packages are for?

Not exactly correct, but good enough: you want 686 if you have a 32 bit
system (that'd include a 32 bit Windows running on 64 bit hardware, I
guess). You want 86_64 if you are on a 64 bit system.

This all assuming you are on Intel or AMD hardware (does Windows run on
anything else?)

regards
- -- t
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)

iEYEARECAAYFAlgbWkcACgkQBcgs9XrR2kYWsQCeNIelbd5hWnUrVUBzk45qhDvj
R4MAn2N89KPYXBtKz0HuOqNCKC3CCCOD
=MEv1
-----END PGP SIGNATURE-----

Re: Windows emacs-25.1 i686 vs x86_64?

[David M. Miller]

On Thursday, November 3, 2016 at 11:32:42 AM UTC-4, David M. Miller wrote:
> When I went to download 25.1 for windows I saw two packages on the main GNU ftp server:
> 
>     emacs-25.1-i686-w64-mingw32.zip
>      - AND -
>     emacs-25.1-x86_64-w64-mingw32.zip
> 
> but no explanation of the difference between them in the README file.
> 
> Also, there seems to be a separate package of "deps" for each:
> 
>     emacs-25-i686-deps.zip
>      - AND -
>     emacs-25-x86_64-deps.zip
> 
> the need for which is also not explained anywhere (dependencies? dependencies for what?).
> 
> Does anyone know what the difference between the i686 and x86_64 packages are?
> Does anyone know what the "deps" packages are for?
> 
> Thanks,
> David

After examining both main emacs packages, I think I figured out the obscure naming conventions: 'i686' refers to the 32-bit version, while 'x86_64' refers to the 64-bit version. So this answers my first question.

I have no intention of doing so, but I still wonder whether the "deps" package is intended for *BUILDING* emacs, or if it provides runtime binary support for some of the more obscure elisp facilities.

Re: Windows emacs-25.1 i686 vs x86_64?

[John Mastro]

> After examining both main emacs packages, I think I figured out the obscure naming conventions: 'i686' refers to the 32-bit version, while 'x86_64' refers to the 64-bit version. So this answers my first question.
> 
> I have no intention of doing so, but I still wonder whether the "deps" package is intended for *BUILDING* emacs, or if it provides runtime binary support for some of the more obscure elisp facilities.

They're runtime dependencies for several optional features. 

        John

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

"David M. Miller" <dmmiller2k@gmail.com> writes:

> On Thursday, November 3, 2016 at 11:32:42 AM UTC-4, David M. Miller wrote:
>> When I went to download 25.1 for windows I saw two packages on the main GNU ftp server:
>> 
>>     emacs-25.1-i686-w64-mingw32.zip
>>      - AND -
>>     emacs-25.1-x86_64-w64-mingw32.zip
>> 
>> but no explanation of the difference between them in the README file.
>> 
>> Also, there seems to be a separate package of "deps" for each:
>> 
>>     emacs-25-i686-deps.zip
>>      - AND -
>>     emacs-25-x86_64-deps.zip
>> 
>> the need for which is also not explained anywhere (dependencies? dependencies for what?).
>> 
>> Does anyone know what the difference between the i686 and x86_64 packages are?
>> Does anyone know what the "deps" packages are for?
>> 
>> Thanks,
>> David
>
> After examining both main emacs packages, I think I figured out the obscure
> naming conventions: 'i686' refers to the 32-bit version, while 'x86_64' refers
> to the 64-bit version. So this answers my first question.
>
> I have no intention of doing so, but I still wonder whether the "deps" package
> is intended for *BUILDING* emacs, or if it provides runtime binary support for
> some of the more obscure elisp facilities.


Apologies for this. I built those packages, and haven't got around to
updating the README. I will try and do so tonight.

i686 and x86_64 are the 32 and 64 bit versions respectively. The names
come from the names for these platforms used in ./configure. Probably
not something that I should be inflicting on users. The "deps" package
are the external dependencies that Emacs uses, for features like XML
parsing, image rendering and so forth. If you unpack these over the
relevant version of Emacs, it's powers will be magically enhanced.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[David M. Miller]

Awesome, thanks. That's what I was asking.

One question though, what runs the shell scripts, absent any recognizable
shell (bash or sh, ash, zsh, etc.) executable in the bin directory?

--David

On Nov 3, 2016 12:28 PM, "Phillip Lord" <phillip.lord@russet.org.uk> wrote:

> "David M. Miller" <dmmiller2k@gmail.com> writes:
>
> > On Thursday, November 3, 2016 at 11:32:42 AM UTC-4, David M. Miller
> wrote:
> >> When I went to download 25.1 for windows I saw two packages on the main
> GNU ftp server:
> >>
> >>     emacs-25.1-i686-w64-mingw32.zip
> >>      - AND -
> >>     emacs-25.1-x86_64-w64-mingw32.zip
> >>
> >> but no explanation of the difference between them in the README file.
> >>
> >> Also, there seems to be a separate package of "deps" for each:
> >>
> >>     emacs-25-i686-deps.zip
> >>      - AND -
> >>     emacs-25-x86_64-deps.zip
> >>
> >> the need for which is also not explained anywhere (dependencies?
> dependencies for what?).
> >>
> >> Does anyone know what the difference between the i686 and x86_64
> packages are?
> >> Does anyone know what the "deps" packages are for?
> >>
> >> Thanks,
> >> David
> >
> > After examining both main emacs packages, I think I figured out the
> obscure
> > naming conventions: 'i686' refers to the 32-bit version, while 'x86_64'
> refers
> > to the 64-bit version. So this answers my first question.
> >
> > I have no intention of doing so, but I still wonder whether the "deps"
> package
> > is intended for *BUILDING* emacs, or if it provides runtime binary
> support for
> > some of the more obscure elisp facilities.
>
>
> Apologies for this. I built those packages, and haven't got around to
> updating the README. I will try and do so tonight.
>
> i686 and x86_64 are the 32 and 64 bit versions respectively. The names
> come from the names for these platforms used in ./configure. Probably
> not something that I should be inflicting on users. The "deps" package
> are the external dependencies that Emacs uses, for features like XML
> parsing, image rendering and so forth. If you unpack these over the
> relevant version of Emacs, it's powers will be magically enhanced.
>
> Phil
>

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

"David M. Miller" <dmmiller2k@gmail.com> writes:

> Awesome, thanks. That's what I was asking.
>
> One question though, what runs the shell scripts, absent any recognizable
> shell (bash or sh, ash, zsh, etc.) executable in the bin directory?


Hopefully nothing!

The files in the dependency packages have been created from MSYS. There
are definately more files in there than necessary.  To be honest, I
thought I had uploaded a more minimal set than this, but apparently not.
I will check.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: phillip.lord@russet.org.uk (Phillip Lord)
> Date: Thu, 03 Nov 2016 16:28:54 +0000
> Cc: help-gnu-emacs@gnu.org
> 
> Apologies for this. I built those packages, and haven't got around to
> updating the README. I will try and do so tonight.

That file is README.W32 in the source tree.  If something needs to be
corrected there, please correct the master copy as well.

Thanks.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: phillip.lord@russet.org.uk (Phillip Lord)
>> Date: Thu, 03 Nov 2016 16:28:54 +0000
>> Cc: help-gnu-emacs@gnu.org
>> 
>> Apologies for this. I built those packages, and haven't got around to
>> updating the README. I will try and do so tonight.
>
> That file is README.W32 in the source tree.  If something needs to be
> corrected there, please correct the master copy as well.


Yes. Do you want to have a look at this patch? The changes are extensive
and include a lot of removals (it's about half the length). I've also
changed the name as .W32 seems liable to confuse in a 64 bit age.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: phillip.lord@russet.org.uk (Phillip Lord)
> Cc: help-gnu-emacs@gnu.org
> Date: Thu, 03 Nov 2016 22:25:16 +0000
> 
> > That file is README.W32 in the source tree.  If something needs to be
> > corrected there, please correct the master copy as well.
> 
> 
> Yes. Do you want to have a look at this patch?

Yes, please.

> I've also changed the name as .W32 seems liable to confuse in a 64
> bit age.

The name of the file is not important, since it's supposed to be
renamed when it's put into the distribution.  So I'd rather not change
the name, because doing so makes investigating repository history
harder.

Thanks.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Eli Zaretskii <eliz@gnu.org> writes:
>> I've also changed the name as .W32 seems liable to confuse in a 64
>> bit age.
>
> The name of the file is not important, since it's supposed to be
> renamed when it's put into the distribution.  So I'd rather not change
> the name, because doing so makes investigating repository history
> harder.

Okay, I'll restore that, and commit to emacs-25.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

[Moving this to emacs-devel.]

> From: phillip.lord@russet.org.uk (Phillip Lord)
> Cc: help-gnu-emacs@gnu.org
> Date: Fri, 04 Nov 2016 10:19:30 +0000
> 
> > The name of the file is not important, since it's supposed to be
> > renamed when it's put into the distribution.  So I'd rather not change
> > the name, because doing so makes investigating repository history
> > harder.
> 
> Okay, I'll restore that, and commit to emacs-25.

I made some minor changes in the file.  Please make sure to leave 2
spaces between sentences, per our conventions (your commit log
messages have the same problem, btw).  Also, please don't remove
information about Windows 9X, as the 32-bit MS-Windows build of Emacs
still supports that.

Thanks.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Eli Zaretskii <eliz@gnu.org> writes:

> [Moving this to emacs-devel.]
>
>> From: phillip.lord@russet.org.uk (Phillip Lord)
>> Cc: help-gnu-emacs@gnu.org
>> Date: Fri, 04 Nov 2016 10:19:30 +0000
>> 
>> > The name of the file is not important, since it's supposed to be
>> > renamed when it's put into the distribution.  So I'd rather not change
>> > the name, because doing so makes investigating repository history
>> > harder.
>> 
>> Okay, I'll restore that, and commit to emacs-25.
>
> I made some minor changes in the file.  Please make sure to leave 2
> spaces between sentences, per our conventions (your commit log
> messages have the same problem, btw).

Yes, indeed, I do forget this often.  It's a convention I have long
dropped in my normal writing if I ever did it.  I will try and remember.

Incidentally, M-q seems to remove the double space.  Am I doing it wrong?

> Also, please don't remove information about Windows 9X, as the 32-bit
> MS-Windows build of Emacs still supports that.

I would argue against this. I removed the material on Windows 9x since
it has long since reached EOL -- 10 years, or 25% of Emacs' existance.
While providing this information somewhere might be useful, having it in
this readme mostly serves to make the readme and to some extent Emacs
appear unmaintained. In addition, it also makes the documentation longer
which adds to the impression that Emacs is hard to use.

I'd be happy to look for somewhere else to put this information, if you
think it need to be retained.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: phillip.lord@russet.org.uk (Phillip Lord)
> Cc: emacs-devel@gnu.org
> Date: Sat, 05 Nov 2016 22:26:38 +0000
> 
> Incidentally, M-q seems to remove the double space.

Not in "emacs -Q", so it's something with your customizations.

> > Also, please don't remove information about Windows 9X, as the 32-bit
> > MS-Windows build of Emacs still supports that.
> 
> I would argue against this. I removed the material on Windows 9x since
> it has long since reached EOL -- 10 years, or 25% of Emacs' existance.

That is true, but we still try supporting those old systems, as they
are widespread in the 3rd world.  We have code whose only purpose is
to continue that support.  We don' take MS EOL decisions as important.

> While providing this information somewhere might be useful, having it in
> this readme mostly serves to make the readme and to some extent Emacs
> appear unmaintained.

I don't see why a document that mentions the latest version of Emacs
could appear unmaintained.  If you'd like we could mention the date of
last update in the file.

> In addition, it also makes the documentation longer which adds to
> the impression that Emacs is hard to use.

One short paragraph is not a significant addition, IMO.

> I'd be happy to look for somewhere else to put this information, if you
> think it need to be retained.

There's no other good place.  This is the file where users should look
for preliminaries for Emacs installation.

Thanks.

Re: Windows emacs-25.1 i686 vs x86_64?

[Paul Eggert]

Eli Zaretskii wrote:
>> > I would argue against this. I removed the material on Windows 9x since
>> > it has long since reached EOL -- 10 years, or 25% of Emacs' existance.
> That is true, but we still try supporting those old systems, as they
> are widespread in the 3rd world.

That may have been true a decade ago, but Windows 9x is no longer a practical 
porting target even in the third world. For what it's worth, statcounter.com 
lists WinME's popularity as dropping from 0.37% of China's desktop market in 
March 2014 to 0.01% in August of this year, and they no longer even bother 
measuring its popularity in Africa.

Besides, at this point it borders on technical malpractice to suggest to users 
that it's OK to use Windows 9x to run Emacs.

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Sun, 6 Nov 2016 01:10:26 -0700
> 
>     That is true, but we still try supporting those old systems, as they
>     are widespread in the 3rd world.
> 
> That may have been true a decade ago, but Windows 9x is no longer a practical porting target even in the third world. For what it's worth, statcounter.com lists WinME's popularity as dropping from 0.37% of China's desktop market in March 2014 to 0.01% in August of this year, and they no longer even bother measuring its popularity in Africa.

Windows/ME was never popular, even when it was released, as it's a
botched OS.  A more interesting statistics is the sum total of all 9X
systems.

Also, even 0.01% in China could be quite a lot of machines.

In any case, the non-removal of support for Windows 9X was an explicit
request from Richard, an opinion which AFAIU he still held 5 years
ago:

  http://lists.gnu.org/archive/html/emacs-devel/2011-07/msg00827.html
  http://lists.gnu.org/archive/html/emacs-devel/2011-07/msg00839.html

> Besides, at this point it borders on technical malpractice to suggest to users that it's OK to use Windows 9x to run Emacs.

How can this be a malpractice, technical or otherwise?  It _is_ okay
to do that for users that don't have access to newer systems.  We are
talking about the same binary, available from the GNU FTP site, that
other users run on the latest Windows systems.  If it's okay on any
other Windows system, it's okay on Windows 9X.

Re: Windows emacs-25.1 i686 vs x86_64?

[Óscar Fuentes]

Eli Zaretskii <eliz@gnu.org> writes:

>> Besides, at this point it borders on technical malpractice to suggest
>> to users that it's OK to use Windows 9x to run Emacs.
>
> How can this be a malpractice, technical or otherwise?

Windows 9X is a virus magnet and reservoir. This is bad for the user of
that machine and for everyone else on the 'net. That's the technical
part. The moral part needs not to be mentioned on this list.

I wholeheartedly agree with Paul on his response.

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: Óscar Fuentes <ofv@wanadoo.es>
> Date: Sun, 06 Nov 2016 17:13:45 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> Besides, at this point it borders on technical malpractice to suggest
> >> to users that it's OK to use Windows 9x to run Emacs.
> >
> > How can this be a malpractice, technical or otherwise?
> 
> Windows 9X is a virus magnet and reservoir.

Which has nothing to do with Emacs, even if it's true.

Re: Windows emacs-25.1 i686 vs x86_64?

[Óscar Fuentes]

Eli Zaretskii <eliz@gnu.org> writes:

>> >> Besides, at this point it borders on technical malpractice to suggest
>> >> to users that it's OK to use Windows 9x to run Emacs.
>> >
>> > How can this be a malpractice, technical or otherwise?
>> 
>> Windows 9X is a virus magnet and reservoir.
>
> Which has nothing to do with Emacs, even if it's true.

As long as Emacs officially supports Windows 9x, it is implicitly
endorsing it. Because the basis of the GNU project is political and
moral, it is important that we send the correct messages to our users.

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: Óscar Fuentes <ofv@wanadoo.es>
> Date: Sun, 06 Nov 2016 18:01:47 +0100
> 
> >> Windows 9X is a virus magnet and reservoir.
> >
> > Which has nothing to do with Emacs, even if it's true.
> 
> As long as Emacs officially supports Windows 9x, it is implicitly
> endorsing it. Because the basis of the GNU project is political and
> moral, it is important that we send the correct messages to our users.

I think you missed something I said here:

  http://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00155.html

Re: Windows emacs-25.1 i686 vs x86_64?

[Óscar Fuentes]

Eli Zaretskii <eliz@gnu.org> writes:

>> >> Windows 9X is a virus magnet and reservoir.
>> >
>> > Which has nothing to do with Emacs, even if it's true.
>> 
>> As long as Emacs officially supports Windows 9x, it is implicitly
>> endorsing it. Because the basis of the GNU project is political and
>> moral, it is important that we send the correct messages to our users.
>
> I think you missed something I said here:
>
>   http://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00155.html

If the "something" is

 If it's okay on any other Windows system, it's okay on Windows 9X.

then that's precisely what I'm disagreeing with.

Having to figure out how to make Emacs work on Windows 9X (*) for
allowing some hypothetical users (**) to keep shooting themselves on the
foot seems like something we could stop doing at this moment.

* Accessing reliable information about which APIs work on Windows 9X is
  not easy at all.

** How many bug reports came from Windows 9X users on the last, say, 8
   years?

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: Óscar Fuentes <ofv@wanadoo.es>
> Date: Sun, 06 Nov 2016 19:26:46 +0100
> 
> >> As long as Emacs officially supports Windows 9x, it is implicitly
> >> endorsing it. Because the basis of the GNU project is political and
> >> moral, it is important that we send the correct messages to our users.
> >
> > I think you missed something I said here:
> >
> >   http://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00155.html
> 
> If the "something" is
> 
>  If it's okay on any other Windows system, it's okay on Windows 9X.

No, it's this:

  In any case, the non-removal of support for Windows 9X was an explicit
  request from Richard, an opinion which AFAIU he still held 5 years
  ago:

    http://lists.gnu.org/archive/html/emacs-devel/2011-07/msg00827.html
    http://lists.gnu.org/archive/html/emacs-devel/2011-07/msg00839.html

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: Óscar Fuentes <ofv@wanadoo.es>
> Date: Sun, 06 Nov 2016 19:26:46 +0100
> 
> Having to figure out how to make Emacs work on Windows 9X (*) for
> allowing some hypothetical users (**) to keep shooting themselves on the
> foot seems like something we could stop doing at this moment.

Just to set the record straight, and to help those who might be
looking for this information:

> * Accessing reliable information about which APIs work on Windows 9X is
>   not easy at all.

One can find that information here:

  http://winapi.freetechsecrets.com/win32/
  http://svn.apache.org/repos/asf/openoffice/trunk/main/sal/systools/win32/uwinapi/unicows.dxp

(The 2nd one lists functions absent from Windows 9X, but added by
UNICOWS.DLL.)

> ** How many bug reports came from Windows 9X users on the last, say, 8
>    years?

Two, AFAIR.  They were both about things that didn't work on 9X, which
we fixed since then.

Re: Windows emacs-25.1 i686 vs x86_64?

[Óscar Fuentes]

Eli Zaretskii <eliz@gnu.org> writes:

>> Having to figure out how to make Emacs work on Windows 9X (*) for
>> allowing some hypothetical users (**) to keep shooting themselves on the
>> foot seems like something we could stop doing at this moment.
>
> Just to set the record straight, and to help those who might be
> looking for this information:
>
>> * Accessing reliable information about which APIs work on Windows 9X is
>>   not easy at all.
>
> One can find that information here:
>
>   http://winapi.freetechsecrets.com/win32/

Hmmm... let's hope it lasts. It is a web version of the help files that
came with Delphi, which IIRC were copyrighted by MS and licensed to
Borland.

>   http://svn.apache.org/repos/asf/openoffice/trunk/main/sal/systools/win32/uwinapi/unicows.dxp
>
> (The 2nd one lists functions absent from Windows 9X, but added by
> UNICOWS.DLL.)

I don't know how a DLL symbol dump can be regarded as documentation but,
anyway, maybe those URLs should be added to the appropriate files in the
Emacs documentation.

>> ** How many bug reports came from Windows 9X users on the last, say, 8
>>    years?
>
> Two, AFAIR.  They were both about things that didn't work on 9X, which
> we fixed since then.

I remember at least one issue raised by a Windows 9X user years ago.
That individual didn't fit the "3rd world" profile at all. IIRC he was
using Windows 9X just because he was change-averse. I always find it
funny that those individuals insist on keeping their old machines with
the correspondingly old OSes while at the same time requesting the
latest versions of third-party software packages :-)

Re: Windows emacs-25.1 i686 vs x86_64?

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > As long as Emacs officially supports Windows 9x, it is implicitly
  > endorsing it. Because the basis of the GNU project is political and
  > moral, it is important that we send the correct messages to our users.

Our ethical objection to Windows applies equally to all versions.
We consider technical issues secondary.

As for security, in principle you can't trust any version of Windows
because it is proprietary software.  The universal back door
was first detected in Windows 98, and is now in Windows 10.
There is no reason to prefer the latter to the former.
Whether versions before Windows 98 had the universal back door,
we don't know.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > As long as Emacs officially supports Windows 9x, it is implicitly
>   > endorsing it. Because the basis of the GNU project is political and
>   > moral, it is important that we send the correct messages to our users.
>
> Our ethical objection to Windows applies equally to all versions.
> We consider technical issues secondary.

If this is true, then in this case, we should be making the decision on
the basis of the technical issue, as the primary issue (that of ethics)
makes no difference here.


Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > > Our ethical objection to Windows applies equally to all versions.
  > > We consider technical issues secondary.

  > If this is true, then in this case, we should be making the decision on
  > the basis of the technical issue, as the primary issue (that of ethics)
  > makes no difference here.

You took one sentence out of context and interpreted it in a way that
is the opposite of my point.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: Windows emacs-25.1 i686 vs x86_64?

[Stefan Monnier]

> In any case, the non-removal of support for Windows 9X was an explicit
> request from Richard, an opinion which AFAIU he still held 5 years
> ago:

5 years is a fairly long time.  AFAIK Windows 98 was released in
... 1998 (surprise surprise), with a second edition in 1999.
So 5 years ago it was 13 years old.  There's a significant different
between 13 years old and 18 years old, so there's a good chance that the
number of users of Window98 is fallen sharply during this time.

Of course, I don't know whether it has, nor whether the intersection of
Windows98 users with Emacs users has fallen sharply.


        Stefan

Re: Windows emacs-25.1 i686 vs x86_64?

[John Wiegley]

>>>>> "EZ" == Eli Zaretskii <eliz@gnu.org> writes:

EZ> In any case, the non-removal of support for Windows 9X was an explicit
EZ> request from Richard, an opinion which AFAIU he still held 5 years ago:

Richard, do you still maintain this position?

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: Windows emacs-25.1 i686 vs x86_64?

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > EZ> In any case, the non-removal of support for Windows 9X was an explicit
  > EZ> request from Richard, an opinion which AFAIU he still held 5 years ago:

  > Richard, do you still maintain this position?

Yes.

We support running our software on various versions of Windows and
MacOS for the sake of users that want to run our programs there, but
that does not mean an endorsement of those systems.  On the contrary,
we say they are unjust and we urge users to stop running them.

We want to teach people to give the highest importance to the ethical
issue, the injustice of nonfree software.  Mere technical issues (such
as state of maintenance) are all secondary, and we want to make our
position on this clear.  So we should not apply words such as
"malpractice" to technical issues.  Windows IS malpractice, and
the worst malpractice is Windows 10.

See http://gnu.org/proprietary/malware-microsoft.html.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: Windows emacs-25.1 i686 vs x86_64?

[Paul Eggert]

Eli Zaretskii wrote:
> Windows/ME was never popular, even when it was released

As I understand it, that "WinME" label is a catchall that includes the 
MS-Windows 9x line, and "0.01%" is so small that I doubt whether it is 
statistically significant. In practice then 9x line has transitioned from 
unsupported (ten years ago) to on-its-way-out (five years ago) to dead (now, 
almost everywhere), and any old decisions that assumed 9x's viability are now 
obsolete.

In GNU projects, we typically stop worrying about an underlying platform when 
its original supplier stops supporting it. For example, Emacs no longer worries 
about IRIX because SGI stopped supporting IRIX in 2013. Although MS-Windows 9x 
is special partly because it was so popular long ago, Emacs need not support 9x 
indefinitely, and Emacs's documentation should not give Emacs users the 
incorrect impression that 9x is still a live platform.

Re: Windows emacs-25.1 i686 vs x86_64?

[Daniel Colascione]

On 11/06/2016 01:54 PM, Paul Eggert wrote:
> Eli Zaretskii wrote:
>> Windows/ME was never popular, even when it was released
>
> As I understand it, that "WinME" label is a catchall that includes the
> MS-Windows 9x line, and "0.01%" is so small that I doubt whether it is
> statistically significant. In practice then 9x line has transitioned
> from unsupported (ten years ago) to on-its-way-out (five years ago) to
> dead (now, almost everywhere), and any old decisions that assumed 9x's
> viability are now obsolete.
>
> In GNU projects, we typically stop worrying about an underlying platform
> when its original supplier stops supporting it. For example, Emacs no
> longer worries about IRIX because SGI stopped supporting IRIX in 2013.
> Although MS-Windows 9x is special partly because it was so popular long
> ago, Emacs need not support 9x indefinitely, and Emacs's documentation
> should not give Emacs users the incorrect impression that 9x is still a
> live platform.

It would be nice to have a data-based condition for dropping 9X support. 
Windows 9X already well beyond the level of dead past which we stop 
supporting other platforms, and it'd be nice to have a clear 
understanding of when we can finally drop 9X support --- this way, we'll 
know when we can drop XP support.

Re: Windows emacs-25.1 i686 vs x86_64?

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > In GNU projects, we typically stop worrying about an underlying
  > platform when its original supplier stops supporting it. For
  > example, Emacs no longer worries about IRIX because SGI stopped
  > supporting IRIX in 2013.

In the GNU Project, the question that matters, for a system version of
no particular importance to us (such as any version of Windows), is
whether users care about that version enough to maintain support for
it.  If they do that, we may as well not delete their code
unless it is getting in the way rather badly.

If we have no direct evidence about whether users care about a certain
version, by default we can suppose that they won't care about a
version that is no longer being maintained.  But that's not the
criterion, just a default way to guess.

Given that Windows is so widely used, and that so many users stick to
old versions of it, it is plausible to me that millions of people
still use Windows 98.  Maybe tens or hundreds of millions.

That number may still be growing.  ISTR that even a few years ago
people were still installing unauthorized copies of Windows 98 on PCs,
because Microsoft made it harder to install subsequent Windows
versions.  It would not surprise me if Windows 98 was installed on
millions of new PCs this year.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: Windows emacs-25.1 i686 vs x86_64?

[Daniel Colascione]

On 11/07/2016 06:43 AM, Richard Stallman wrote:
> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > In GNU projects, we typically stop worrying about an underlying
>   > platform when its original supplier stops supporting it. For
>   > example, Emacs no longer worries about IRIX because SGI stopped
>   > supporting IRIX in 2013.
>
> In the GNU Project, the question that matters, for a system version of
> no particular importance to us (such as any version of Windows), is
> whether users care about that version enough to maintain support for
> it.  If they do that, we may as well not delete their code
> unless it is getting in the way rather badly.
>
> If we have no direct evidence about whether users care about a certain
> version, by default we can suppose that they won't care about a
> version that is no longer being maintained.  But that's not the
> criterion, just a default way to guess.
>
> Given that Windows is so widely used, and that so many users stick to
> old versions of it, it is plausible to me that millions of people
> still use Windows 98.  Maybe tens or hundreds of millions.
>
> That number may still be growing.  ISTR that even a few years ago
> people were still installing unauthorized copies of Windows 98 on PCs,
> because Microsoft made it harder to install subsequent Windows
> versions.  It would not surprise me if Windows 98 was installed on
> millions of new PCs this year.
>

What data would, in principle, convince you otherwise?

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > In GNU projects, we typically stop worrying about an underlying
>   > platform when its original supplier stops supporting it. For
>   > example, Emacs no longer worries about IRIX because SGI stopped
>   > supporting IRIX in 2013.
>
> In the GNU Project, the question that matters, for a system version of
> no particular importance to us (such as any version of Windows), is
> whether users care about that version enough to maintain support for
> it.  If they do that, we may as well not delete their code
> unless it is getting in the way rather badly.


The original issue that came up here is not about code but about
documentation. My desire is to have the readme a) as short as possible
and b) not make Emacs appear unmaintained.


> If we have no direct evidence about whether users care about a certain
> version, by default we can suppose that they won't care about a
> version that is no longer being maintained.  But that's not the
> criterion, just a default way to guess.
>
> Given that Windows is so widely used, and that so many users stick to
> old versions of it, it is plausible to me that millions of people
> still use Windows 98.  Maybe tens or hundreds of millions.
>
> That number may still be growing.  ISTR that even a few years ago
> people were still installing unauthorized copies of Windows 98 on PCs,
> because Microsoft made it harder to install subsequent Windows
> versions.  It would not surprise me if Windows 98 was installed on
> millions of new PCs this year.


All this may be true, but it does not help. We have no direct evidence
about whether users care about a certain version. Unless we have a way
of gaining that evidence, we are left with supposition.

What we do know is that we get very few bug reports for Windows 98/95.
The only thing that I could think of further to get more information
would be to check the download logs; does ftp.gnu.org store user agent
strings or equivalent. This would at least give us some evidence about
Windows 9x -- also Windows XP.

Incidentally, even if Emacs is supported on these platforms, I have
not tested the binary downloads on any of them, nor do I have the
capability to do so.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > What we do know is that we get very few bug reports for Windows 98/95.

That suggests we don't have many users on those systems.

We could put in code in the next release to display a message,
"If you use GNU Emacs on Windows 95 or Windows 98,
please send email to <some email address>.  We are thinking
of deleting the support for those systems in a future Emacs version."

  > The only thing that I could think of further to get more information
  > would be to check the download logs; does ftp.gnu.org store user agent
  > strings or equivalent. This would at least give us some evidence about
  > Windows 9x -- also Windows XP.

That is a good idea.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > What we do know is that we get very few bug reports for Windows 98/95.
>
> That suggests we don't have many users on those systems.
>
> We could put in code in the next release to display a message,
> "If you use GNU Emacs on Windows 95 or Windows 98,
> please send email to <some email address>.  We are thinking
> of deleting the support for those systems in a future Emacs version."

We don't need "If you use..." -- Emacs should know OS it is using.
I am happy to add this to Emacs-26 (perhaps as an addition to the splash
screen). For "<some email address>" we could use a debbugs email. This
would then provide us a generalized mechanism for deprecation of OS
support. The only difficulty I can see is finding a copy of Win 98 on
which to test this.

Can I suggest we add 9x, and Windows XP into the list for Emacs-26?

>   > The only thing that I could think of further to get more information
>   > would be to check the download logs; does ftp.gnu.org store user agent
>   > strings or equivalent. This would at least give us some evidence about
>   > Windows 9x -- also Windows XP.
>
> That is a good idea.

I will investigate and see if this data is available.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: phillip.lord@russet.org.uk (Phillip Lord)
> Cc: eggert@cs.ucla.edu,  eliz@gnu.org,  emacs-devel@gnu.org
> Date: Tue, 08 Nov 2016 14:52:44 +0000
> 
> Can I suggest we add 9x, and Windows XP into the list for Emacs-26?

My main development machine runs Windows XP, so if you want to drop
support for it, you should ask me to resign first.

Btw, there are 2 versions of Windows between 9X and XP, which we still
support, and you didn't mention.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: phillip.lord@russet.org.uk (Phillip Lord)
>> Cc: eggert@cs.ucla.edu,  eliz@gnu.org,  emacs-devel@gnu.org
>> Date: Tue, 08 Nov 2016 14:52:44 +0000
>> 
>> Can I suggest we add 9x, and Windows XP into the list for Emacs-26?
>
> My main development machine runs Windows XP, so if you want to drop
> support for it, you should ask me to resign first.

Yes, I remember you saying and I remain surprised that you use such an
old operating system, and presumably computer. I guess we wait till your
machine breaks before we move to XP.

> Btw, there are 2 versions of Windows between 9X and XP, which we still
> support, and you didn't mention.

Good point, I had forgotten, my memory of those times is hazy.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: phillip.lord@russet.org.uk (Phillip Lord)
> Cc: rms@gnu.org,  eggert@cs.ucla.edu,  emacs-devel@gnu.org
> Date: Wed, 09 Nov 2016 16:50:24 +0000
> 
> > My main development machine runs Windows XP, so if you want to drop
> > support for it, you should ask me to resign first.
> 
> Yes, I remember you saying and I remain surprised that you use such an
> old operating system, and presumably computer. I guess we wait till your
> machine breaks before we move to XP.

My machine is just 4-year old, so it is unlikely to break any time
soon (barring force majeure).

> > Btw, there are 2 versions of Windows between 9X and XP, which we still
> > support, and you didn't mention.
> 
> Good point, I had forgotten, my memory of those times is hazy.

FWIW, I think considering deprecation of XP, or any other of the OS
versions from the NT family, is unjustified, because, unlike Windows
9X, the number of features that Emacs needs missing from those older
NT-family versions is very small, while all the significant features
we want -- Unicode APIs, file security and access control, Uniscribe
complex script shaping, etc. -- are present in all of those versions.
So removing support for these versions will require some non-trivial
amount of work, and will risk introducing bugs, while the gains will
be insignificant to non-existent.  By contrast, testing for
availability of a function is just a few boilerplate lines of code.
It really isn't worth the hassle.

Re: Windows emacs-25.1 i686 vs x86_64?

[Stefan Monnier]

> Yes, I remember you saying and I remain surprised that you use such an
> old operating system, and presumably computer. I guess we wait till your
> machine breaks before we move to XP.

FWIW, I don't see much point in discontinuing support for XP.  At least
not until we bump into some situation where preserving compatibility
with XP causes significant extra work, or until XP is sufficiently rare.
Currently XP is pretty far from rare in my experience (many users stuck
to XP and resisted upgrading because the subsequent versions delivered
by Microsoft were perceived to be worse in some respects, which is why
Microsoft ended up having to extend support of XP until as late as 2014).


        Stefan

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> Yes, I remember you saying and I remain surprised that you use such an
>> old operating system, and presumably computer. I guess we wait till your
>> machine breaks before we move to XP.
>
> FWIW, I don't see much point in discontinuing support for XP.  At least
> not until we bump into some situation where preserving compatibility
> with XP causes significant extra work, or until XP is sufficiently rare.
> Currently XP is pretty far from rare in my experience (many users stuck
> to XP and resisted upgrading because the subsequent versions delivered
> by Microsoft were perceived to be worse in some respects, which is why
> Microsoft ended up having to extend support of XP until as late as 2014).


Yes, that's all true. I hung onto XP until the bitter end. And, it
doesn't require any statements in the README which is what got me here
in the first place.

No worries.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[martin rudalics]

 > This
 > would then provide us a generalized mechanism for deprecation of OS
 > support. The only difficulty I can see is finding a copy of Win 98 on
 > which to test this.
 >
 > Can I suggest we add 9x, and Windows XP into the list for Emacs-26?

Why deprecate support for systems that is not broken?  Shouldn't we
rather concentrate our efforts on fixing support for systems like Cairo
or Gnustep?  Would continuing support for Windows 98 hinder us in any
way doing that?

martin

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

phillip.lord@russet.org.uk (Phillip Lord) writes:
> We don't need "If you use..." -- Emacs should know OS it is using.
> I am happy to add this to Emacs-26 (perhaps as an addition to the splash
> screen). For "<some email address>" we could use a debbugs email. This
> would then provide us a generalized mechanism for deprecation of OS
> support. The only difficulty I can see is finding a copy of Win 98 on
> which to test this.
>
> Can I suggest we add 9x, and Windows XP into the list for Emacs-26?
>
>>   > The only thing that I could think of further to get more information
>>   > would be to check the download logs; does ftp.gnu.org store user agent
>>   > strings or equivalent. This would at least give us some evidence about
>>   > Windows 9x -- also Windows XP.
>>
>> That is a good idea.
>
> I will investigate and see if this data is available.


So, I managed to get info about one day -- there have been 32,000
accesses to ftp.gnu.org. None (for any project, let alone emacs) have
come with a Win 9x user agent.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Paul Eggert]

On 11/07/2016 06:43 AM, Richard Stallman wrote:
> even a few years ago
> people were still installing unauthorized copies of Windows 98 on PCs,
> because Microsoft made it harder to install subsequent Windows
> versions.

That was true fifteen years ago, but unauthorized versions of 
more-recent MS-Windows systems have been widely available for ages and 
they are the ones invariably installed nowadays. Last year, for example, 
Microsoft announced[1] that it would give free Windows 10 licenses to 
users running unauthorized copies of Windows 7 or 8.1 because 
unauthorized versions were so popular (they reportedly had most of the 
Chinese market). StatCounter reports that in China, where unauthorized 
use is common, Windows XP had over 90% of desktop use by 2008 (with the 
remainder mostly being later versions of Windows), and that nowadays 
Windows 7 and Windows 10 are both more popular than XP, with the 9x line 
being too small to count separately during this period.

So in practice Windows 9x is dead, and covering Windows 9x prominently 
in Emacs documentation wastes users' time and makes us look technically 
dated.

1. Hachman M. Microsoft offers amnesty of sorts to pirates of older 
Windows software. PCWorld 2015-10-29. 
http://www.pcworld.com/article/2999443/windows/microsoft-offers-amnesty-of-sorts-to-pirates-of-older-windows-software.html

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> Cc: eliz@gnu.org, emacs-devel@gnu.org, phillip.lord@russet.org.uk
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Mon, 7 Nov 2016 08:34:48 -0800
> 
> covering Windows 9x prominently in Emacs documentation wastes users'
> time and makes us look technically dated.

A short paragraph in a README is not "prominent documentation".  Let's
not lose the perspective when trying to make a point, okay?

Re: Windows emacs-25.1 i686 vs x86_64?

[Perry E. Metzger]

On Mon, 07 Nov 2016 20:07:13 +0200 Eli Zaretskii <eliz@gnu.org> wrote:
> > covering Windows 9x prominently in Emacs documentation wastes
> > users' time and makes us look technically dated.
> 
> A short paragraph in a README is not "prominent documentation".
> Let's not lose the perspective when trying to make a point, okay?

I'm somewhat surprised that we're discussing support for a proprietary
OS that has been off vendor extended support for over a decade and
which no one uses any more.

I have to assume the main reason for that is that some people
(including RMS) were under the mistaken impression that it is still
widely used, which it is not. Data has already been provided to back
that up.

(If you buy a new computer in China it comes with Windows 10, though
of course not necessarily a legal copy. No one would willingly buy a
machine with Windows 98 installed. It doesn't even support modern
hardware.)

Perry
-- 
Perry E. Metzger		perry@piermont.com

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> Date: Mon, 7 Nov 2016 14:02:28 -0500
> From: "Perry E. Metzger" <perry@piermont.com>
> Cc: Paul Eggert <eggert@cs.ucla.edu>, phillip.lord@russet.org.uk,
>  rms@gnu.org, emacs-devel@gnu.org
> 
> I'm somewhat surprised that we're discussing support for a proprietary
> OS that has been off vendor extended support for over a decade and
> which no one uses any more.

We aren't.  "Discussing support" would be talking about some code that
needs to be written or rewritten or changed in support of that OS.
Nothing of the sort is on the table; the code is already written, was
written long ago, actually, and no one suggested to work on it, let
alone improve it.

What we are doing is we are bikeshedding about a short paragraph of
text that I added back to a README, that's all.  It's a mouse that
gave birth to a mountain.  I guess we have nothing more important to
do with our time.

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Eli Zaretskii <eliz@gnu.org> writes:
> What we are doing is we are bikeshedding about a short paragraph of
> text that I added back to a README, that's all.  It's a mouse that
> gave birth to a mountain.  I guess we have nothing more important to
> do with our time.

My motivation for raising the issue and deleting that text was not one
of bikeshedding.

Of course, as a piece of software with a 40 year history, Emacs is both
bound to and probably should be a relatively conservative project. But,
at the same time, it does need to change and adapt with the times, at
many different levels. And, it has done: dash.el at a code level, a
package system at adminstrative level and now double buffering in the
display.

Emacs is old. Because of this it displays in part maturity and in parts
senescence. Keep the former is a good thing, the latter is not. What
concerns me, here, is that in a document which might be the first thing
that a prospective new user sees, we are saying "if you use Windows 9x".

Every year, I get the privilege of teaching some of the next generation
of programmers, and I live program with Emacs in front of them. Windows
98 became obsolete when they were eight and was released before they
were born. We're never going to be the coolest kids on the block and to
try would be Dad-dancing. But, appearing to live in the past is not good
either (cue Jethro Tull).

Here I leave the discussion.

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Stefan Monnier]

> (If you buy a new computer in China it comes with Windows 10, though
> of course not necessarily a legal copy. No one would willingly buy a
> machine with Windows 98 installed. It doesn't even support modern
> hardware.)

The argument was not for people using new machines with old OS, but
people using old machines (with accompanying old OS).


        Stefan

Re: Windows emacs-25.1 i686 vs x86_64?

[Perry E. Metzger]

On Mon, 07 Nov 2016 14:39:26 -0500 Stefan Monnier
<monnier@iro.umontreal.ca> wrote:
> > (If you buy a new computer in China it comes with Windows 10,
> > though of course not necessarily a legal copy. No one would
> > willingly buy a machine with Windows 98 installed. It doesn't
> > even support modern hardware.)  
> 
> The argument was not for people using new machines with old OS, but
> people using old machines (with accompanying old OS).

Sure, but there's very little evidence that many of those are extant.

Perry
-- 
Perry E. Metzger		perry@piermont.com

Re: Windows emacs-25.1 i686 vs x86_64?

[Stefan Monnier]

> Sure, but there's very little evidence that many of those are extant.

Five years ago, I would have wholeheartedly agreed with you.  And yet,
at that point, some user complained about something that broke under
Windows98 and in their context it was still in wide use and they went
through a fair bit of trouble helping us debug this use-case until it
worked again.

So my gut feeling still agrees with yours, but I wouldn't necessarily
trust it.


        Stefan

Re: Windows emacs-25.1 i686 vs x86_64?

[Perry E. Metzger]

On Mon, 07 Nov 2016 15:10:57 -0500 Stefan Monnier
<monnier@iro.umontreal.ca> wrote:
> > Sure, but there's very little evidence that many of those are
> > extant.  
> 
> Five years ago, I would have wholeheartedly agreed with you.  And
> yet, at that point, some user complained about something that broke
> under Windows98 and in their context it was still in wide use and
> they went through a fair bit of trouble helping us debug this
> use-case until it worked again.
> 
> So my gut feeling still agrees with yours, but I wouldn't
> necessarily trust it.

I think deliberately breaking it seems like it would be antisocial,
but going through substantial trouble (that is, say, holding back
some sort of improved functionality) to make sure it keeps working
also seems unreasonable.

Perry
-- 
Perry E. Metzger		perry@piermont.com

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> Date: Mon, 7 Nov 2016 15:22:28 -0500
> From: "Perry E. Metzger" <perry@piermont.com>
> Cc: emacs-devel@gnu.org
> 
> I think deliberately breaking it seems like it would be antisocial,
> but going through substantial trouble (that is, say, holding back
> some sort of improved functionality) to make sure it keeps working
> also seems unreasonable.

We never do the latter.  If some functionality cannot work on old
Windows systems, there's a runtime test which disables the feature on
those systems, with some suitable return value or error message.

This has been the Emacs practice for years.

Re: Windows emacs-25.1 i686 vs x86_64?

[Elias Mårtenson]

[-- Attachment #1: Type: text/plain, Size: 1259 bytes --]

On 8 November 2016 at 04:27, Eli Zaretskii <eliz@gnu.org> wrote:

> > Date: Mon, 7 Nov 2016 15:22:28 -0500
> > From: "Perry E. Metzger" <perry@piermont.com>
> > Cc: emacs-devel@gnu.org
> >
> > I think deliberately breaking it seems like it would be antisocial,
> > but going through substantial trouble (that is, say, holding back
> > some sort of improved functionality) to make sure it keeps working
> > also seems unreasonable.
>
> We never do the latter.  If some functionality cannot work on old
> Windows systems, there's a runtime test which disables the feature on
> those systems, with some suitable return value or error message.
>
> This has been the Emacs practice for years.
>

FWIW, my impression after reading this thread wasn't merely about the text
in the README, but rather as to whether time should be spent even
considering Windows 9x when working on Emacs.

I took the liberty to search the archives, and found several instances this
year alone where time was spent discussing whether or not to use one
function or another because they weren't supposed on these older versions
of Windows.

At the risk of putting opinions into other people's mouths, I do think that
those are the kinds of discussions no one really wants.

Regards,
Elias

[-- Attachment #2: Type: text/html, Size: 1870 bytes --]

Re: Windows emacs-25.1 i686 vs x86_64?

[Eli Zaretskii]

> From: Elias Mårtenson <lokedhs@gmail.com>
> Date: Tue, 8 Nov 2016 11:48:18 +0800
> Cc: "Perry E. Metzger" <perry@piermont.com>, Stefan Monnier <monnier@iro.umontreal.ca>, 
> 	emacs-devel <emacs-devel@gnu.org>
> 
>     I think deliberately breaking it seems like it would be antisocial,
>     > but going through substantial trouble (that is, say, holding back
>     > some sort of improved functionality) to make sure it keeps working
>     > also seems unreasonable.
> 
>     We never do the latter.  If some functionality cannot work on old
>     Windows systems, there's a runtime test which disables the feature on
>     those systems, with some suitable return value or error message.
> 
>     This has been the Emacs practice for years.
> 
> FWIW, my impression after reading this thread wasn't merely about the text in the README, but rather as to whether time should be spent even considering Windows 9x when working on Emacs.

It started because I asked not to delete that text.  Leaving that text
alone would exactly mean we don't need to spend any time even
considering Windows 9X, whereas this discussion does require us to
consider it.

> I took the liberty to search the archives, and found several instances this year alone where time was spent discussing whether or not to use one function or another because they weren't supposed on these older versions of Windows.

During this year, I see just one discussion (in January) of a certain
patch wrt how to adapt it to some older systems, and how to fix bugs
and issues revealed on those systems (including, but not limited to,
9X).

> At the risk of putting opinions into other people's mouths, I do think that those are the kinds of discussions no one really wants.

Some of those discussions (usually, comments to patches) cannot be
avoided, because some library functions and APIs aren't available on
all OS versions.  Failure to either use more widely available APIs or
provide a run-time test for their availability will lead to an Emacs
binary that will refuse to start on some versions of Windows, even
though the offending API is not needed by that user in that session.
The result will be that Emacs can only be trusted to run on the system
where it was built.

On Unix, these tests are done at configure time, and therefore the
produced binary cannot be safely copied to another system.  By
contrast, on Windows, it is very customary for users to download
binaries compiled on some other system, so configure-time testing
cannot be used, and must be replaced with run-time testing, if and
when the corresponding APIs are needed for some Emacs feature and
alternative APIs don't exist.  People who contribute code to Emacs
aren't always aware of this issue, so it comes up in discussing
patches.  The code for these tests is boilerplate, but it must be
there for each API that is not guaranteed to exist on all supported
versions of the OS.

And of course, this isn't limited to Windows 9X, since each new
version of Windows introduces APIs that aren't available in previous
versions.  So providing bleeding-edge features in Emacs on Windows
will always need to include run-time tests for availability of the
required new APIs, because we do want Emacs to continue being able to
run on older systems, even if those bleeding-edge features might not
be available there.  We already have a few features that are disabled,
or fall back on simple replacements, on versions of Windows newer than
9X, sometimes much newer (e.g., creation and resolution of symbolic
links aren't supported below Vista).

Re: Windows emacs-25.1 i686 vs x86_64?

[Perry E. Metzger]

On Mon, 07 Nov 2016 09:43:23 -0500 Richard Stallman <rms@gnu.org>
wrote:
> Given that Windows is so widely used, and that so many users stick
> to old versions of it, it is plausible to me that millions of people
> still use Windows 98.  Maybe tens or hundreds of millions.

It is not likely. There is statistical information that was
cited earlier in this thread. Windows 9x and ME are now below 0.01%
usage. How much below is hard to measure.

Emacs was likely never particularly popular with people willing to
hold on to an OS for that long either, as such people are typically
not technical users (and in fact, the users of those systems are in
fact typically not people at all in the sense that such systems,
when they remain, are embedded systems doing things like balancing
car wheels and running old cash registers).

So, given a fraction of a percent userbase and a fraction of a
percent for that OS, I think it would be fairly safe to assume we're
talking about a fraction of a percent of a fraction of a percent,
probably too small to measure.

Such people also are unlikely to have very good ability to
use the internet because their machines would be taken over by
malware within seconds of attaching to an open network, and because
no web browsers exist for such platforms capable of using modern web
sites. I would guess, therefore, that they're unlikely to be
downloading new copies of Emacs as they arise.

There is nothing, IMHO, wrong with telling such users that they will
have to use Emacs 25.1 and before and documenting that. I would be
quite sincerely surprised, however, if anyone at all was
inconvenienced.

> That number may still be growing.  ISTR that even a few years ago
> people were still installing unauthorized copies of Windows 98 on
> PCs, because Microsoft made it harder to install subsequent Windows
> versions.

That has not been true in a long time. As noted, the statistics do not
bear this out, and in fact, it is not hard to install a more recent
version of Windows even without Microsoft permission. If you're in a
country like China you can get your hands on unlocked much more recent
versions of Windows without paying for it.

Indeed, the situation is such that Microsoft more or less made it
possible to install Windows 10 without any license key. The only real
effect of this is that you get a watermark in one corner of the
screen telling you that you should register your copy.

> It would not surprise me if Windows 98 was installed on
> millions of new PCs this year.

I will bet you this is not the case.

Perry
-- 
Perry E. Metzger		perry@piermont.com

Re: Windows emacs-25.1 i686 vs x86_64?

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > It is not likely. There is statistical information that was
  > cited earlier in this thread. Windows 9x and ME are now below 0.01%
  > usage. How much below is hard to measure.

It seems I was mistaken in what I remembered.  Maybe it was Windows XP
that was still being installed a few years ago.

-- 
Dr Richard Stallman
President, Free Software Foundation (gnu.org, fsf.org)
Internet Hall-of-Famer (internethalloffame.org)
Skype: No way! See stallman.org/skype.html.

Re: Windows emacs-25.1 i686 vs x86_64?

[Perry E. Metzger]

On Tue, 08 Nov 2016 08:55:23 -0500 Richard Stallman <rms@gnu.org>
wrote:
>   > It is not likely. There is statistical information that was
>   > cited earlier in this thread. Windows 9x and ME are now below
>   > 0.01% usage. How much below is hard to measure.
> 
> It seems I was mistaken in what I remembered.  Maybe it was Windows
> XP that was still being installed a few years ago.

Windows XP is, sadly, still used by some people, though a large
fraction of that use is in embedded systems (like older medical
equipment and old ATMs) rather than on desktops. I don't think anyone
is installing it into new things terribly much if at all, but yes,
there's a certain amount of it out there.

Official end of support from Microsoft happened on April 8, 2014. I
would say that in a few years it will not be of much importance, but
for now there's still a *bit*.


Perry
-- 
Perry E. Metzger		perry@piermont.com

Re: Windows emacs-25.1 i686 vs x86_64?

[Phillip Lord]

Eli Zaretskii <eliz@gnu.org> writes:
>> > Also, please don't remove information about Windows 9X, as the 32-bit
>> > MS-Windows build of Emacs still supports that.
>> 
>> I would argue against this. I removed the material on Windows 9x since
>> it has long since reached EOL -- 10 years, or 25% of Emacs' existance.
>
> That is true, but we still try supporting those old systems, as they
> are widespread in the 3rd world.  We have code whose only purpose is
> to continue that support.  We don' take MS EOL decisions as important.

EOL decisions are important. Of course, this does not mean that you have
to take them as gospel, but to ignore them is not a good course, I
think.

Still I see others have taken up this issue in the thread, so I will not
revisit the issue of whether 95/98 are supported further, and so only
comment on it's position in the documentation.

>> While providing this information somewhere might be useful, having it in
>> this readme mostly serves to make the readme and to some extent Emacs
>> appear unmaintained.
>
> I don't see why a document that mentions the latest version of Emacs
> could appear unmaintained.  If you'd like we could mention the date of
> last update in the file.


I think that this is not the issue -- in general, if I read a document
that has lots of advice on what to do with long obsolete operating
systems, I would just assume that it has not been maintained.


>> In addition, it also makes the documentation longer which adds to
>> the impression that Emacs is hard to use.
>
> One short paragraph is not a significant addition, IMO.


The file is already very long for a readme. My experience of watching
new Emacs users is that there experience is very much "do I have to read
all of this to edit a file".

Reducing things like this readme to an absolute minimum seems a useful
aim. How to support a very, very old operating system seems a poor thing
to include. To steal a quote "It's a beautiful thing, the Destruction of
words."

Incidentally, on this topic, is this paragraph (adding in 2001) still at
all relevant?

  Virus scanners

  Some virus scanners interfere with Emacs' use of subprocesses.  If you
  are unable to use subprocesses and you use Dr. Solomon's WinGuard or
  McAfee's Vshield, turn off "Scan all files" (WinGuard) or "boot sector
  scanning" (McAfee exclusion properties).


>> I'd be happy to look for somewhere else to put this information, if you
>> think it need to be retained.
>
> There's no other good place.  This is the file where users should look
> for preliminaries for Emacs installation.


README.W32.obsolete?

Phil

Re: Windows emacs-25.1 i686 vs x86_64?

[Noam Postavsky]

On Sat, Nov 5, 2016 at 6:26 PM, Phillip Lord <phillip.lord@russet.org.uk> wrote:
>> I made some minor changes in the file.  Please make sure to leave 2
>> spaces between sentences, per our conventions (your commit log
>> messages have the same problem, btw).
>
> Yes, indeed, I do forget this often.  It's a convention I have long
> dropped in my normal writing if I ever did it.  I will try and remember.
>
> Incidentally, M-q seems to remove the double space.  Am I doing it wrong?

That probably means you've set sentence-end-double-space to nil. The
Emacs repo has a .dir-locals.el file that sets it to t, but if you're
using magit it currently won't apply when writing commit messages, see
https://github.com/magit/magit/issues/2848