bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Drew Adams]

(To be sure, I've downloaded the zip file twice now, and extracted it.)

I got it from here: http://gnu.mirrors.pair.com/gnu/emacs/windows/

This is the zip file I picked up:

emacs-25.1-2-x86_64-w64-mingw32.zip     29-Nov-2016 14:12   91M

But when I open Emacs and use M-x emacs-version it tells me that it is
Emacs 21.1.1.  And you can see that below, as well.

What gives?  Is this correct or a bug?  How are users supposed to
understand the version numbering - is this 25.1.1 or 25.1.2 or something
else?

And if this is not 25.1.2, where can users get 25.1.2?

In GNU Emacs 25.1.1 (x86_64-w64-mingw32)
 of 2016-11-15
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --without-dbus --without-compress-install 'CFLAGS=-O2
 -static -g3''

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Noam Postavsky]

On Thu, Jan 26, 2017 at 4:00 PM, Drew Adams <drew.adams@oracle.com> wrote:
> (To be sure, I've downloaded the zip file twice now, and extracted it.)
>
> I got it from here: http://gnu.mirrors.pair.com/gnu/emacs/windows/
>
> This is the zip file I picked up:
>
> emacs-25.1-2-x86_64-w64-mingw32.zip     29-Nov-2016 14:12   91M
>
> But when I open Emacs and use M-x emacs-version it tells me that it is
> Emacs 21.1.1.  And you can see that below, as well.
>
> What gives?  Is this correct or a bug?  How are users supposed to
> understand the version numbering - is this 25.1.1 or 25.1.2 or something
> else?

It's 25.1-2, that is a second release of version 25.1. The first one
was compiled without optimization.

https://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00293.html

>
> And if this is not 25.1.2, where can users get 25.1.2?
>
> In GNU Emacs 25.1.1 (x86_64-w64-mingw32)
>  of 2016-11-15
> Windowing system distributor `Microsoft Corp.', version 6.1.7601
> Configured using:
>  `configure --without-dbus --without-compress-install 'CFLAGS=-O2
>  -static -g3''
>
>
>

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Drew Adams]

> > This is the zip file I picked up:
> > emacs-25.1-2-x86_64-w64-mingw32.zip     29-Nov-2016 14:12   91M
> > But when I open Emacs and use M-x emacs-version it tells me that it is
> > Emacs 21.1.1.  And you can see that below, as well.
> >
> > What gives?  Is this correct or a bug?  How are users supposed to
> > understand the version numbering - is this 25.1.1 or 25.1.2 or
> > something else?
> 
> It's 25.1-2, that is a second release of version 25.1. The first one
> was compiled without optimization.
> https://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00293.html

Sorry, but I still don't understand.  What's the difference between
25.1-2 and 25.1.2?  Do both 25.1-1 (whatever/wherever that is) and
25.1-2 correspond to 25.1.1?

Where does a user see the version number 25.1-2 (or whatever kind of
number it is, if it is not a version number)?  How, from Emacs itself,
does a user tell the difference between 25.1-1 and 25.1-2?

These were the zip files downloaded, respectively:

 emacs-25.1-x86_64-mingw32.zip
 emacs-25.1-2-x86_64-w64-mingw32.zip

What does a user (or code) look for within Emacs, to tell the difference?

And are these releases?  Release candidates?  Pretest builds?  How to know?

Is all this supposed to be clear to users?  Where do they learn about
such nomenclature (zip and release)?  At the mirrors where you download
these things there is no description/explanation.  (They don't even tell
you what the difference is between zips x86 and i86.)

I realize that building and uploading Emacs for Windows is a volunteer
effort, and I'm grateful for it.  But can we perhaps make the zip naming
clearer?  (I'm guessing at this point that there might not be any
Emacs-detectable version difference, in spite of the zip name difference.)

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[npostavs]

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

> These were the zip files downloaded, respectively:
>
>  emacs-25.1-x86_64-mingw32.zip
>  emacs-25.1-2-x86_64-w64-mingw32.zip
>
> What does a user (or code) look for within Emacs, to tell the difference?

25.1-2 should run faster, otherwise it's identical to 25.1.

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Eli Zaretskii]

> Date: Thu, 26 Jan 2017 14:30:58 -0800 (PST)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: 25548@debbugs.gnu.org
> 
> > It's 25.1-2, that is a second release of version 25.1. The first one
> > was compiled without optimization.
> > https://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00293.html
> 
> Sorry, but I still don't understand.  What's the difference between
> 25.1-2 and 25.1.2?

They are builds of the same codebase with different compiler options.

> Do both 25.1-1 (whatever/wherever that is) and 25.1-2 correspond to
> 25.1.1?

They both correspond to 25.1.  There's no 25.1.1 release; 25.1.1 is
build #1 of the 25.1 release.

> Where does a user see the version number 25.1-2 (or whatever kind of
> number it is, if it is not a version number)?

It's not a version number.  It's a name of the zip file, and was give
"-2" to distinguish it from the original upload.

> How, from Emacs itself, does a user tell the difference between
> 25.1-1 and 25.1-2?

The value of system-configuration-options should be different.

> These were the zip files downloaded, respectively:
> 
>  emacs-25.1-x86_64-mingw32.zip
>  emacs-25.1-2-x86_64-w64-mingw32.zip
> 
> What does a user (or code) look for within Emacs, to tell the difference?

See above.

> And are these releases?  Release candidates?  Pretest builds?

These are releases.  Pretests have a 4-field versions, like
25.1.91.2.  Release candidates have "RC" in the name of the zip file.

> How to know?

See the node "Version Info" in the ELisp manual.

> Is all this supposed to be clear to users?

This information is not necessarily of importance for users.  It is
important in bug reports, which is why report-emacs-bug includes it.
But inquisitive users can find the information in the Emacs manuals.

> At the mirrors where you download these things there is no
> description/explanation.  (They don't even tell you what the
> difference is between zips x86 and i86.)

The explanations you are looking for are in the README file that is
right near the zip archives.

> I realize that building and uploading Emacs for Windows is a volunteer
> effort, and I'm grateful for it.  But can we perhaps make the zip naming
> clearer?

Did you read the README file?

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Drew Adams]

> > > It's 25.1-2, that is a second release of version 25.1. The first one
> > > was compiled without optimization.
> > > https://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00293.html
> >
> > Sorry, but I still don't understand.  What's the difference between
> > 25.1-2 and 25.1.2?
> 
> They are builds of the same codebase with different compiler options.

Really?  And just where is build 25.1.2 (that's a 2, not a 1)?
Does it exist somewhere?  I have 25.1-2, but you say that 25.1.2
has different compiler options from 25.1-2.

> > Do both 25.1-1 (whatever/wherever that is) and 25.1-2 correspond to
> > 25.1.1?
> 
> They both correspond to 25.1.  There's no 25.1.1 release; 25.1.1 is
> build #1 of the 25.1 release.

So presumably 25.1.2 should be build #2 of 25.1.  Where is 25.1.2?
Why 25.1-2 instead of 25.1.2?

> > Where does a user see the version number 25.1-2 (or whatever kind of
> > number it is, if it is not a version number)?
> 
> It's not a version number.  It's a name of the zip file, and was give
> "-2" to distinguish it from the original upload.

You just said it is also the build number - build #2 of 25.1.

> > How, from Emacs itself, does a user tell the difference between
> > 25.1-1 and 25.1-2?
> 
> The value of system-configuration-options should be different.

Show does a user or Lisp code see those options from _within_
Emacs?  Where is that information given to users?

> > These were the zip files downloaded, respectively:
> >
> >  emacs-25.1-x86_64-mingw32.zip
> >  emacs-25.1-2-x86_64-w64-mingw32.zip
> >
> > What does a user (or code) look for within Emacs, to tell the
> > difference?
> 
> See above.

See above.  _Within_ Emacs.  We have Lisp functions to
determine versions, but builds?  What are the functions
for that?

> > And are these releases?  Release candidates?  Pretest builds?
> 
> These are releases.  Pretests have a 4-field versions, like
> 25.1.91.2.  Release candidates have "RC" in the name of the
> zip file.

You said these are builds of the same release, _not_ releases.
You said, I think, that 25.1-2 and 25.1.2 are the same thing,
and they are the second build of release 25.1.

> > How to know?
> 
> See the node "Version Info" in the ELisp manual.

The only thing there (that I can see) that remotely refers
to something like "25.1.2" is the undescribed, unexplained
_output_ of `(emacs-version)':

"GNU Emacs 24.5.1 (x86_64-unknown-linux-gnu, GTK+ Version 3.16)
           ^^^^^^
        of 2015-06-01"

Where in that node is a description of what "25.1.2" or
"25.1-2" means - e.g., that it means build #2 of release 25.1?

> > Is all this supposed to be clear to users?
> 
> This information is not necessarily of importance for users.  It is
> important in bug reports, which is why report-emacs-bug includes it.
> But inquisitive users can find the information in the Emacs manuals.

I haven't seen where.  I see such a label in passing, in the
output of `emacs-version', but it is unexplained, AFAICT.

> > At the mirrors where you download these things there is no
> > description/explanation.  (They don't even tell you what the
> > difference is between zips x86 and i86.)
> 
> The explanations you are looking for are in the README file that is
> right near the zip archives.

I don't think so.  That README is similar to (but also different
from) the README.W32 that is included in the zip file.  AFAICT,
neither file describes the 25.1-2 (or whatever other number)
that appears in the zip-file name.  Searching for "zip" in each
of those readme files does not find anything that talks about this. 

> > I realize that building and uploading Emacs for Windows is a volunteer
> > effort, and I'm grateful for it.  But can we perhaps make the zip naming
> > clearer?
> 
> Did you read the README file?

I've read them both, and searched them for "zip".  Have you taken
a look yourself, in light of the question of this bug report?
Can you point me to the passage that explains the 25.1-2 in the
zip name, which I've not been able to find?  Thx.

Another question is why we have two different designations
for build 2 of release 25.1: "25.1.2" and "25.1-2".  I think
you said that those represent the same thing.  If so, why
use two different representations, Occam wonders?  And if
not, what is the difference between them?

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Eli Zaretskii]

> Date: Fri, 27 Jan 2017 07:40:41 -0800 (PST)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: npostavs@users.sourceforge.net, 25548@debbugs.gnu.org
> 
> > > > It's 25.1-2, that is a second release of version 25.1. The first one
> > > > was compiled without optimization.
> > > > https://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00293.html
> > >
> > > Sorry, but I still don't understand.  What's the difference between
> > > 25.1-2 and 25.1.2?
> > 
> > They are builds of the same codebase with different compiler options.
> 
> Really?  And just where is build 25.1.2 (that's a 2, not a 1)?

It doesn't exist.  I thought you were asking about the difference
between 25.1.1 and 25.1-2.

> > They both correspond to 25.1.  There's no 25.1.1 release; 25.1.1 is
> > build #1 of the 25.1 release.
> 
> So presumably 25.1.2 should be build #2 of 25.1.  Where is 25.1.2?
> Why 25.1-2 instead of 25.1.2?

Presumably because Phillip made a full bootstrap, which starts
counting builds anew.

> > > How, from Emacs itself, does a user tell the difference between
> > > 25.1-1 and 25.1-2?
> > 
> > The value of system-configuration-options should be different.
> 
> Show does a user or Lisp code see those options from _within_
> Emacs?

With "M-: system-configuration-options RET".

> > > How to know?
> > 
> > See the node "Version Info" in the ELisp manual.
> 
> The only thing there (that I can see) that remotely refers
> to something like "25.1.2" is the undescribed, unexplained
> _output_ of `(emacs-version)':
> 
> "GNU Emacs 24.5.1 (x86_64-unknown-linux-gnu, GTK+ Version 3.16)
>            ^^^^^^
>         of 2015-06-01"

No, it also explains how to distinguish official releases from pretest
versions.

> > > Is all this supposed to be clear to users?
> > 
> > This information is not necessarily of importance for users.  It is
> > important in bug reports, which is why report-emacs-bug includes it.
> > But inquisitive users can find the information in the Emacs manuals.
> 
> I haven't seen where.

I've told you where.

> > > At the mirrors where you download these things there is no
> > > description/explanation.  (They don't even tell you what the
> > > difference is between zips x86 and i86.)
> > 
> > The explanations you are looking for are in the README file that is
> > right near the zip archives.
> 
> I don't think so.  That README is similar to (but also different
> from) the README.W32 that is included in the zip file.  AFAICT,
> neither file describes the 25.1-2 (or whatever other number)
> that appears in the zip-file name.  Searching for "zip" in each
> of those readme files does not find anything that talks about this. 

It also says this:

  There are two binary distributions named
  emacs-n-x86_64-w64-mingw32.zip and
  macs-n-i686-w64-mingw32.zip. These are for 64-bit and 32-bit
  machines respectively.

You asked about 32-bit vs the 64-bit versions; that part answers the
question.

> > > I realize that building and uploading Emacs for Windows is a volunteer
> > > effort, and I'm grateful for it.  But can we perhaps make the zip naming
> > > clearer?
> > 
> > Did you read the README file?
> 
> I've read them both, and searched them for "zip".  Have you taken
> a look yourself, in light of the question of this bug report?
> Can you point me to the passage that explains the 25.1-2 in the
> zip name, which I've not been able to find?

That's the "-n-" part above.

> Another question is why we have two different designations
> for build 2 of release 25.1: "25.1.2" and "25.1-2".

No, we have only one: 25.1.2.  The other one is just the name of a
file.

> I think you said that those represent the same thing.  If so, why
> use two different representations, Occam wonders?

Who or what is Occam?  I thought I was talking to Drew.

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Drew Adams]

> > > > How, from Emacs itself, does a user tell the difference between
> > > > 25.1-1 and 25.1-2?
> > >
> > > The value of system-configuration-options should be different.
> >
> > Show does a user or Lisp code see those options from _within_
> > Emacs?
> 
> With "M-: system-configuration-options RET".

I see this from `C-h v system-configuration-options":
"--without-dbus --without-compress-install 'CFLAGS=-O2 -static -g3'"

How does that tell you anything about the label "25.1.1",
which appears in the output of `M-x emacs-version' or
the label "25.1-2" that appears in the zip-file name?

> > > > How to know?
> > >
> > > See the node "Version Info" in the ELisp manual.
> >
> > The only thing there (that I can see) that remotely refers
> > to something like "25.1.2" is the undescribed, unexplained
> > _output_ of `(emacs-version)':
> >
> > "GNU Emacs 24.5.1 (x86_64-unknown-linux-gnu, GTK+ Version 3.16)
> >            ^^^^^^
> >         of 2015-06-01"
> 
> No, it also explains how to distinguish official releases from pretest
> versions.

That's not what I said.  It tells you nothing about what
"24.5.1" (or in the current case, "25.1.1") means in the
output of `emacs-version'.  And it tells you nothing about
what (in the current case) "25.1-2" means in the name of
the zip file.

> > > > Is all this supposed to be clear to users?
> > >
> > > This information is not necessarily of importance for users.  It is
> > > important in bug reports, which is why report-emacs-bug includes it.
> > > But inquisitive users can find the information in the Emacs manuals.
> >
> > I haven't seen where.
> 
> I've told you where.

No, I don't think you have.  See above.  I see no explanation
of the "25.1.1" nomenclature that appears in the output of
`emacs-version'.  And I see no explanation of the "25.1-2"
nomenclature that appears in the zip-file name.
 
> > > > At the mirrors where you download these things there is no
> > > > description/explanation.  (They don't even tell you what the
> > > > difference is between zips x86 and i86.)
> > >
> > > The explanations you are looking for are in the README file that is
> > > right near the zip archives.
> >
> > I don't think so.  That README is similar to (but also different
> > from) the README.W32 that is included in the zip file.  AFAICT,
> > neither file describes the 25.1-2 (or whatever other number)
> > that appears in the zip-file name.  Searching for "zip" in each
> > of those readme files does not find anything that talks about this.
> 
> It also says this:
> 
>   There are two binary distributions named
>   emacs-n-x86_64-w64-mingw32.zip and
>   macs-n-i686-w64-mingw32.zip. These are for 64-bit and 32-bit
>   machines respectively.
>
> You asked about 32-bit vs the 64-bit versions; that part
> answers the question.

Yes, but it does not answer the question about "25.1-2".

> > Can you point me to the passage that explains the 25.1-2 in the
> > zip name, which I've not been able to find?
> 
> That's the "-n-" part above.

Yes, n = 25.1-2, here.  But where is the explanation of it,
saying what n = 25.1-2 means?

> > Another question is why we have two different designations
> > for build 2 of release 25.1: "25.1.2" and "25.1-2".
> 
> No, we have only one: 25.1.2.  The other one is just the name of a
> file.

I think you (and I) meant 25.1.1, no?  It is apparently 25.1.1
that corresponds to 25.1-2 in the file name.  When I evaluate
`(emacs-version)' in the build from the zip file with name
25.1-2 it tells me:

GNU Emacs 25.1.1 (x86_64-w64-mingw32) of 2016-11-15

It says 25.1.1, not 25.1.2.

Still looking for an explanation of such numbering, in both
the output of `emacs-version' and the zip-file name.

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Noam Postavsky]

On Fri, Jan 27, 2017 at 1:13 PM, Drew Adams <drew.adams@oracle.com> wrote:
>
> GNU Emacs 25.1.1 (x86_64-w64-mingw32) of 2016-11-15
>
> It says 25.1.1, not 25.1.2.
>
> Still looking for an explanation of such numbering, in both
> the output of `emacs-version' and the zip-file name.

The zip-file name convention of X.Y-2 is new, and it hasn't been
documented yet. Adding a mention in the README next to the archives
would make sense.

Regarding 25.1.1 vs 25.1.2 I think this text in `(elisp) Version Info'
explains it:

 -- Variable: emacs-version
     The value of this variable is the version of Emacs being run.  It
     is a string such as ‘"23.1.1"’.  The last number in this string is
     not really part of the Emacs release version number; it is
     incremented each time Emacs is built in any given directory.

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Drew Adams]

> The zip-file name convention of X.Y-2 is new, and it hasn't been
> documented yet. Adding a mention in the README next to the archives
> would make sense.

Thank you.

bug#25548: 25.1; Wrong Emacs version for Emacs 25.1.2 (?)

[Eli Zaretskii]

> Date: Fri, 27 Jan 2017 10:13:04 -0800 (PST)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: npostavs@users.sourceforge.net, 25548@debbugs.gnu.org
> 
> > > Show does a user or Lisp code see those options from _within_
> > > Emacs?
> > 
> > With "M-: system-configuration-options RET".
> 
> I see this from `C-h v system-configuration-options":
> "--without-dbus --without-compress-install 'CFLAGS=-O2 -static -g3'"
> 
> How does that tell you anything about the label "25.1.1",
> which appears in the output of `M-x emacs-version' or
> the label "25.1-2" that appears in the zip-file name?

It tells you what the version number means.

> > > > See the node "Version Info" in the ELisp manual.
> > >
> > > The only thing there (that I can see) that remotely refers
> > > to something like "25.1.2" is the undescribed, unexplained
> > > _output_ of `(emacs-version)':
> > >
> > > "GNU Emacs 24.5.1 (x86_64-unknown-linux-gnu, GTK+ Version 3.16)
> > >            ^^^^^^
> > >         of 2015-06-01"
> > 
> > No, it also explains how to distinguish official releases from pretest
> > versions.
> 
> That's not what I said.  It tells you nothing about what
> "24.5.1" (or in the current case, "25.1.1") means in the
> output of `emacs-version'.  And it tells you nothing about
> what (in the current case) "25.1-2" means in the name of
> the zip file.

It does to me.

> > > Can you point me to the passage that explains the 25.1-2 in the
> > > zip name, which I've not been able to find?
> > 
> > That's the "-n-" part above.
> 
> Yes, n = 25.1-2, here.  But where is the explanation of it,
> saying what n = 25.1-2 means?

The explanation is in the value of system-configuration-options.

> GNU Emacs 25.1.1 (x86_64-w64-mingw32) of 2016-11-15
> 
> It says 25.1.1, not 25.1.2.

Yes.

> Still looking for an explanation of such numbering, in both
> the output of `emacs-version' and the zip-file name.

Time to finish this nonsense.