Emacs pretest 22.0.99

Emacs pretest 22.0.99

[Chong Yidong]

The Emacs 22.0.99 pretest is now available at

ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta

Note that python-mode has been removed from this tarball.  The legal
status of python.el is currently being discussed; see the thread on
emacs-pretest-bug.  Maybe it will be resolved in time for Emacs 22.1,
maybe not.  In the meantime, the version of python.el maintained by
Dave Love can be found at

http://www.emacswiki.org/cgi-bin/wiki/PythonMode

If no problems are found in this over the next few days, I think we
should release Emacs 22.1; but I guess that's up to RMS to decide.

Re: Emacs pretest 22.0.99

[Nick Roberts]

 > If no problems are found in this over the next few days, I think we
 > should release Emacs 22.1; but I guess that's up to RMS to decide.

This sounds sensible to me.  How about numbering bug fix releases, which
presumably could be made shortly after the main release if necessary, 22.1.1,
22.1.2, ... this time?  In the past there has been some confusion about the
content of 21.2, 21.3 and 21.4, which AFAICS were bascically just respins of
21.1.

If signifacnt changes are released before Unicode Emacs, these could be
numbered 22.2, 22.3 etc.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> From: Nick Roberts <nickrob@snap.net.nz>
> Date: Tue, 24 Apr 2007 17:48:56 +1200
> Cc: emacs-devel@gnu.org
> 
> How about numbering bug fix releases, which
> presumably could be made shortly after the main release if necessary, 22.1.1,
> 22.1.2, ... this time?

This was discussed at length some time ago, but I believe we didn't
decide to change the numbering.  It cannot be changed without changes
elsewhere; e.g., report-emacs-bug knows about the numbering scheme,
and so do the files involved in setting the minor version number
during the build process.

In other words, I think it's too late to change this for Emacs 22.

Re: Emacs pretest 22.0.99

[Nick Roberts]

 > > How about numbering bug fix releases, which presumably could be made
 > > shortly after the main release if necessary, 22.1.1, 22.1.2, ... this
 > > time?
 > 
 > This was discussed at length some time ago, but I believe we didn't
 > decide to change the numbering.

Many schemes were discussed, I can't remember them all but I'm sure you're
right.  Perhaps we could decide now.

 >                                  It cannot be changed without changes
 > elsewhere; e.g., report-emacs-bug knows about the numbering scheme,
 > and so do the files involved in setting the minor version number
 > during the build process.
 >
 > In other words, I think it's too late to change this for Emacs 22.

Well report-emacs-bug just seems to look for three dots:

(string-match "\\..*\\..*\\." "22.0.98.1")
2
(string-match "\\..*\\..*\\." "22.1.1")
nil
(string-match "\\..*\\..*\\." "22.1")
nil

What are the other files?  Do they just test for three dots also?

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> From: Nick Roberts <nickrob@snap.net.nz>
> Date: Tue, 24 Apr 2007 18:26:39 +1200
> Cc: cyd@stupidchicken.com, emacs-devel@gnu.org
> 
>  >                                  It cannot be changed without changes
>  > elsewhere; e.g., report-emacs-bug knows about the numbering scheme,
>  > and so do the files involved in setting the minor version number
>  > during the build process.
>  >
>  > In other words, I think it's too late to change this for Emacs 22.
> 
> Well report-emacs-bug just seems to look for three dots:
> 
> (string-match "\\..*\\..*\\." "22.0.98.1")

Yes, and you seemed to be suggesting that we change that, so that now
non-pretest builds are X.Y.Z.1,2,3,... while pretest versions are
A.B.C.D.E.1,2,3...

> What are the other files?

loadup.el, for example, and make-dist.  Probably others as well, I
don't remember.

> Do they just test for three dots also?

Some do, some don't, please take a look.

Re: Emacs pretest 22.0.99

[Nick Roberts]

 > > Well report-emacs-bug just seems to look for three dots:
 > > 
 > > (string-match "\\..*\\..*\\." "22.0.98.1")
 > 
 > Yes, and you seemed to be suggesting that we change that, so that now
 > non-pretest builds are X.Y.Z.1,2,3,... while pretest versions are
 > A.B.C.D.E.1,2,3...

It looks like 21.1 is actually 21.1.1,  21.2 is actually 21.2.1 etc - I didn't
realise that.  No, I was suggesting 22.1.2, 22.1.3 etc as bug-fix release
numbers (now omitting 22.1.1).

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> From: Nick Roberts <nickrob@snap.net.nz>
> Date: Wed, 25 Apr 2007 10:53:31 +1200
> Cc: cyd@stupidchicken.com, emacs-devel@gnu.org
> 
> It looks like 21.1 is actually 21.1.1,  21.2 is actually 21.2.1 etc - I didn't
> realise that.  No, I was suggesting 22.1.2, 22.1.3 etc as bug-fix release
> numbers (now omitting 22.1.1).

If we release version 22.1.2, then its first build on the end-user
machine would be 22.1.2.1, the second build 22.1.2.2, etc.  This is
what I meant: you are suggesting to enlarge the number of digits in
the version string.

Re: Emacs pretest 22.0.99

[Nick Roberts]

 > > It looks like 21.1 is actually 21.1.1, 21.2 is actually 21.2.1 etc - I
 > > didn't realise that.  No, I was suggesting 22.1.2, 22.1.3 etc as bug-fix
 > > release numbers (now omitting 22.1.1).
 > 
 > If we release version 22.1.2, then its first build on the end-user
 > machine would be 22.1.2.1, the second build 22.1.2.2, etc.  This is
 > what I meant: you are suggesting to enlarge the number of digits in
 > the version string.

I think I see what you are saying: if someone builds the release 22.1 more than
once it's built as 22.1.1, 22.1.2 etc. - I've only built from CVS.

OK, sorry for my confusion.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> From: Nick Roberts <nickrob@snap.net.nz>
> Date: Wed, 25 Apr 2007 15:57:38 +1200
> Cc: cyd@stupidchicken.com, emacs-devel@gnu.org
> 
> I think I see what you are saying: if someone builds the release 22.1 more than
> once it's built as 22.1.1, 22.1.2 etc.

Yes, that's right.

Re: Emacs pretest 22.0.99

[Richard Stallman]

I do not want to change the way we number Emacs versions.
Please let's not have another discussion of it.

Re: Emacs pretest 22.0.99

[Nick Roberts]

 > I do not want to change the way we number Emacs versions.
 > Please let's not have another discussion of it.

I don't think we discussed numbering bug-fix releases, but if your mind is
closed to this, that's another matter.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Emacs pretest 22.0.99

[Richard Stallman]

    I don't think we discussed numbering bug-fix releases, but if your mind is
    closed to this, that's another matter.

I recall we discussed several schemes for numbering that covered all
sorts of releases.

Re: Emacs pretest 22.0.99

[Lennart Borgman (gmail)]

Chong Yidong wrote:
> The Emacs 22.0.99 pretest is now available at
> 
> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta
> 
> Note that python-mode has been removed from this tarball.  The legal
> status of python.el is currently being discussed; see the thread on
> emacs-pretest-bug.  Maybe it will be resolved in time for Emacs 22.1,
> maybe not.  In the meantime, the version of python.el maintained by
> Dave Love can be found at
> 
> http://www.emacswiki.org/cgi-bin/wiki/PythonMode
> 
> If no problems are found in this over the next few days, I think we
> should release Emacs 22.1; but I guess that's up to RMS to decide.


I think the version number for w32 was not changed in emacs/nt/emacs.rc

Re: Emacs pretest 22.0.99

[Lennart Borgman (gmail)]

Lennart Borgman (gmail) wrote:
> Chong Yidong wrote:
>> The Emacs 22.0.99 pretest is now available at
>>
>> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
>> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta
>>
>> Note that python-mode has been removed from this tarball.  The legal
>> status of python.el is currently being discussed; see the thread on
>> emacs-pretest-bug.  Maybe it will be resolved in time for Emacs 22.1,
>> maybe not.  In the meantime, the version of python.el maintained by
>> Dave Love can be found at
>>
>> http://www.emacswiki.org/cgi-bin/wiki/PythonMode
>>
>> If no problems are found in this over the next few days, I think we
>> should release Emacs 22.1; but I guess that's up to RMS to decide.
> 
> 
> I think the version number for w32 was not changed in emacs/nt/emacs.rc


I have uploaded new (unpatched) binaries, but the executables is marked 
as 0.98 as per above.

Re: Emacs pretest 22.0.99

[Jason Rumney]

Lennart Borgman (gmail) wrote:
> I think the version number for w32 was not changed in emacs/nt/emacs.rc

Did you build from the tarball that Chong built?

Re: Emacs pretest 22.0.99

[Lennart Borgman (gmail)]

Jason Rumney wrote:
> Lennart Borgman (gmail) wrote:
>> I think the version number for w32 was not changed in emacs/nt/emacs.rc
> 
> Did you build from the tarball that Chong built?

No, from the CVS, but does not Chong make the tarball directly from the CVS?

Re: Emacs pretest 22.0.99

[Jason Rumney]

Lennart Borgman (gmail) wrote:
> Jason Rumney wrote:
>> Lennart Borgman (gmail) wrote:
>>> I think the version number for w32 was not changed in emacs/nt/emacs.rc
>>
>> Did you build from the tarball that Chong built?
>
> No, from the CVS, but does not Chong make the tarball directly from 
> the CVS?

That file was changed in the EMACS_22_BASE branch of CVS too, and is 
tagged as EMACS_PRETEST_22_0_99.

Did you just checkout HEAD and build from there? If so, you have built 
the wrong sources.

Re: Emacs pretest 22.0.99

[Lennart Borgman (gmail)]

Jason Rumney wrote:
> Lennart Borgman (gmail) wrote:
>> Jason Rumney wrote:
>>> Lennart Borgman (gmail) wrote:
>>>> I think the version number for w32 was not changed in emacs/nt/emacs.rc
>>>
>>> Did you build from the tarball that Chong built?
>>
>> No, from the CVS, but does not Chong make the tarball directly from 
>> the CVS?
> 
> That file was changed in the EMACS_22_BASE branch of CVS too, and is 
> tagged as EMACS_PRETEST_22_0_99.
> 
> Did you just checkout HEAD and build from there? If so, you have built 
> the wrong sources.


Yes, that is what I did. Can you please tell me what I need to do now 
then? I know nothing about CVS branching.

Re: Emacs pretest 22.0.99

[Jason Rumney]

Lennart Borgman (gmail) wrote:
> Yes, that is what I did. Can you please tell me what I need to do now 
> then? I know nothing about CVS branching.

Read the message "Emacs 22 branch created.", or better, build from the 
official tarballs.

Re: Emacs pretest 22.0.99

[Lennart Borgman (gmail)]

Jason Rumney wrote:
> Lennart Borgman (gmail) wrote:
>> Yes, that is what I did. Can you please tell me what I need to do now 
>> then? I know nothing about CVS branching.
> 
> Read the message "Emacs 22 branch created.", or better, build from the 
> official tarballs.

Thanks, that helped. I have uploaded new binaries for w32:

   http://ourcomments.org/cgi-bin/emacsw32-dl-latest.pl

Re: Emacs pretest 22.0.99

[Takashi Hiromatsu]

At Mon, 23 Apr 2007 22:51:56 -0400,
'Chong Yidong' wrote:
> 
> The Emacs 22.0.99 pretest is now available at
> 
> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta

I succeeded to build w32 version by Cygwin latest distribution except make
version. My make is version 3.80.

But after some modification in lisp directory, 'make recompile' failed.

In makefile, variable EMACS is set as
    EMACS = $(THISDIR)/../bin/emacs.exe
but, it should be set as
    EMACS = $(THISDIR)/../src/oo-spd/i386/temacs.exe
for 'make recompile', isn't it?

Takashi Hiromatsu

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> Date: Tue, 24 Apr 2007 17:03:27 +0900
> From: Takashi Hiromatsu <matsuan@ca2.so-net.ne.jp>
> Cc: emacs-devel@gnu.org
> 
> At Mon, 23 Apr 2007 22:51:56 -0400,
> 'Chong Yidong' wrote:
> > 
> > The Emacs 22.0.99 pretest is now available at
> > 
> > ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
> > ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta
> 
> I succeeded to build w32 version by Cygwin latest distribution except make
> version. My make is version 3.80.

Thank you for your report.

Is this a Cygwin build or a native Windows build?  That is, is the
resulting binary a Cygwin program or a native Windows programs?

If the latter, then using Cygwin Make is generally considered to not
work (see nt/INSTALL for details).

> But after some modification in lisp directory, 'make recompile' failed.
> 
> In makefile, variable EMACS is set as
>     EMACS = $(THISDIR)/../bin/emacs.exe
> but, it should be set as
>     EMACS = $(THISDIR)/../src/oo-spd/i386/temacs.exe
> for 'make recompile', isn't it?

No, the default is okay for the case that you didn't change any
sources since the last build, and assuming that Emacs binary is
installed in-place in the `bin' subdirectory of the top-level Emacs
directory.  If you have emacs.exe in some other place, you need to
override the value of EMACS from the command line, for example:

   make recompile EMACS=../src/oo-spd/i386/emacs.exe

In any case, there's no need to use temacs.exe to recompile Lisp
files, you can use emacs.exe in the src/oo-spd directory.

Re: Emacs pretest 22.0.99

[Takashi Hiromatsu]

[-- Attachment #1: Type: text/plain, Size: 2427 bytes --]

> > > ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
> > > ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta
> > 
> > I succeeded to build w32 version by Cygwin latest distribution except make
> > version. My make is version 3.80.
> 
> Thank you for your report.
> 
> Is this a Cygwin build or a native Windows build?  That is, is the
> resulting binary a Cygwin program or a native Windows programs?
> 
> If the latter, then using Cygwin Make is generally considered to not
> work (see nt/INSTALL for details).
Off course, my build is for native Windows.

I konw nt/INSTALL.
|    cygwin compiled make 3.80:           fails?[6]     fails?[6]
|    cygwin compiled make 3.81:           fails         fails?[6]
|    mingw32 compiled make 3.79.1:        okay          okay
|    mingw32 compiled make 3.80:          okay          okay[6]
|    mingw32 compiled make 3.81:          okay          okay[7]
|
|  Notes:
|
|    [1] doesn't cope with makefiles with DOS line endings, so must mount
|        emacs source with text!=binary.
|    [2] fails when needs to invoke shell commands; okay invoking gcc etc.
|    [3] requires LC_MESSAGES support to build; cannot build with early
|        versions of cygwin.
|    [4] may fail on Windows 9X and Windows ME; if so, install Bash.
|    [5] fails when building leim due to the use of cygwin style paths.
|        May work if building emacs without leim.
|    [6] not recommended; please report if you try this combination.
Then I reported that I succeeded with "cygwin compiled make 3.80:"

Off course, before configuration, I had to make un-comment 3 lines in
nt/gmage.defs like as attached patch. Then the binary, I bulld, works well.

> > In makefile, variable EMACS is set as
> >     EMACS = $(THISDIR)/../bin/emacs.exe
> > but, it should be set as
> >     EMACS = $(THISDIR)/../src/oo-spd/i386/temacs.exe
> > for 'make recompile', isn't it?
> 
> No, the default is okay for the case that you didn't change any
> sources since the last build, and assuming that Emacs binary is
> installed in-place in the `bin' subdirectory of the top-level Emacs
> directory.  If you have emacs.exe in some other place, you need to
> override the value of EMACS from the command line, for example:
> 
>    make recompile EMACS=../src/oo-spd/i386/emacs.exe
Thank you for your explanation. I hope that nt/INSTALL have some sentences
like you expalined.

Takashi Hiromatsu


[-- Attachment #2: cygwin.patch --]
[-- Type: application/octet-stream, Size: 475 bytes --]

--- nt/gmake.defs.orgi	2006-08-25 13:42:35.474498400 +0900
+++ nt/gmake.defs	2006-08-25 13:42:42.234353400 +0900
@@ -84,9 +84,9 @@
 # remedied before putting this cygpath kludge back in.
 
 # Convert CURDIR to native file name, if in Cygwin format
-#ifeq "$(shell cygpath $(CURDIR))" "$(CURDIR)"
-#CURDIR	        := $(shell cygpath -m $(CURDIR))
-#endif
+ifeq "$(shell cygpath $(CURDIR))" "$(CURDIR)"
+CURDIR	        := $(shell cygpath -m $(CURDIR))
+endif
 
 THISDIR		= .
 

[-- Attachment #3: Type: text/plain, Size: 1 bytes --]



[-- Attachment #4: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> Date: Wed, 25 Apr 2007 09:41:45 +0900
> From: Takashi Hiromatsu <matsuan@ca2.so-net.ne.jp>
> 
> |    cygwin compiled make 3.80:           fails?[6]     fails?[6]
> |    cygwin compiled make 3.81:           fails         fails?[6]
> |    mingw32 compiled make 3.79.1:        okay          okay
> |    mingw32 compiled make 3.80:          okay          okay[6]
> |    mingw32 compiled make 3.81:          okay          okay[7]
> |
> |  Notes:
> |
> |    [1] doesn't cope with makefiles with DOS line endings, so must mount
> |        emacs source with text!=binary.
> |    [2] fails when needs to invoke shell commands; okay invoking gcc etc.
> |    [3] requires LC_MESSAGES support to build; cannot build with early
> |        versions of cygwin.
> |    [4] may fail on Windows 9X and Windows ME; if so, install Bash.
> |    [5] fails when building leim due to the use of cygwin style paths.
> |        May work if building emacs without leim.
> |    [6] not recommended; please report if you try this combination.
> Then I reported that I succeeded with "cygwin compiled make 3.80:"

Was that with or without a sh.exe on your PATH?  If you do have sh.exe
on PATH, then what kind of sh.exe is it? is it the Cygwin port of Bash
or something else?

> Thank you for your explanation. I hope that nt/INSTALL have some sentences
> like you expalined.

I will see what I can do, thanks.

Re: Emacs pretest 22.0.99

[Takashi Hiromatsu]

At Fri, 27 Apr 2007 16:21:37 +0300,
Eli Zaretskii wrote:
> 
> > Date: Wed, 25 Apr 2007 09:41:45 +0900
> > From: Takashi Hiromatsu <matsuan@ca2.so-net.ne.jp>
> > 
> > |    cygwin compiled make 3.80:           fails?[6]     fails?[6]
> > |    cygwin compiled make 3.81:           fails         fails?[6]
> > |    mingw32 compiled make 3.79.1:        okay          okay
> > |    mingw32 compiled make 3.80:          okay          okay[6]
> > |    mingw32 compiled make 3.81:          okay          okay[7]
> > |
> > |  Notes:
> > |
> > |    [1] doesn't cope with makefiles with DOS line endings, so must mount
> > |        emacs source with text!=binary.
> > |    [2] fails when needs to invoke shell commands; okay invoking gcc etc.
> > |    [3] requires LC_MESSAGES support to build; cannot build with early
> > |        versions of cygwin.
> > |    [4] may fail on Windows 9X and Windows ME; if so, install Bash.
> > |    [5] fails when building leim due to the use of cygwin style paths.
> > |        May work if building emacs without leim.
> > |    [6] not recommended; please report if you try this combination.
> > Then I reported that I succeeded with "cygwin compiled make 3.80:"
> 
> Was that with or without a sh.exe on your PATH?  If you do have sh.exe
> on PATH, then what kind of sh.exe is it? is it the Cygwin port of Bash
> or something else?
I used Cygwin bash for build.

I will explain how I build NTEmacs on Cygwin.
----------------------------------------------------------------------------

First, I installed Cygwin distribution fully. (not a real latest)
My Cygwin make version is 3.80. (latest is 3.81)

Second, I got header files and "dll"s, that I need, from
gnuwin32(http://gnuwin32.sourceforge.net/packages.html).

bash$ mkdir /src/emacs
bash$ mkdir /src/emacs/dll

Then, I put all the required files, header files and "dll"s, in the
/src/emacs/dll directory.

bash$ cd /src/emacs
bash$ wget ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
bash$ tar zxvf emacs-22.0.99.tar.gz

Then I made un-comment 3 lines in emacs-22.0.99/nt/gmage.defs.

-#ifeq "$(shell cygpath $(CURDIR))" "$(CURDIR)"
-#CURDIR	        := $(shell cygpath -m $(CURDIR))
-#endif
+ifeq "$(shell cygpath $(CURDIR))" "$(CURDIR)"
+CURDIR	        := $(shell cygpath -m $(CURDIR))
+endif

bash$ cd emacs-22.0.99/nt
bash$ ./configure.bat --with-gcc --no-cygwin --no-debug \
--prefix d:/cygwin/usr/local/emacs/22.0.99 \
--cflags -I/src/emacs/cvs/dll
bash$ make all
bash$ make install
bash$ cp /src/emacs/dll/*.dll /usr/local/emacs/22.0.99/bin
bash$ /usr/local/emacs/22.0.99/bin/emacs.exe &

Then emacs start to work very well.

Regards,

Takashi Hiromatsu

Re: Emacs pretest 22.0.99

[Eli Zaretskii]

> Date: Tue, 01 May 2007 16:08:19 +0900
> From: Takashi Hiromatsu <matsuan@ca2.so-net.ne.jp>
> Cc: Eli Zaretskii <eliz@gnu.org>
> 
> I will explain how I build NTEmacs on Cygwin.

Thanks, I added the necessary information to nt/INSTALL.

Re: Emacs pretest 22.0.99

[Giorgos Keramidas]

On 2007-04-23 22:51, Chong Yidong <cyd@stupidchicken.com> wrote:
> The Emacs 22.0.99 pretest is now available at
>
> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.99.tar.gz
> ftp://alpha.gnu.org/gnu/emacs/pretest/emacs-22.0.98-22.0.99.xdelta

Tested successfully on FreeBSD/i386 on:
FreeBSD 7.0-CURRENT #0: Wed Apr 25 01:36:57 EEST 2007

I've updated the editors/emacs-devel port of FreeBSD so until
a new tarball is released we'll be using 22.0.99 now ;)

Regards,
Giorgos