Re: Failure building Emacs master

Re: Failure building Emacs master

[Angelo Graziosi]

> If so, the attached patch should implement your suggestion.


It seems the patch is not yet applied on master..

  Angelo

Re: Failure building Emacs master

[Paul Eggert]

On 03/21/2017 09:45 AM, Angelo Graziosi wrote:
> It seems the patch is not yet applied on master.. 

I applied it just now (with streamlined diagnostics), since you and I 
liked the idea and there wasn't any opposing opinion. We can always 
change the default back if there are problems.

Re: Failure building Emacs master

[Eli Zaretskii]

> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Tue, 21 Mar 2017 11:24:23 -0700
> Cc: Emacs developers <emacs-devel@gnu.org>
> 
> On 03/21/2017 09:45 AM, Angelo Graziosi wrote:
> > It seems the patch is not yet applied on master.. 
> 
> I applied it just now (with streamlined diagnostics), since you and I 
> liked the idea and there wasn't any opposing opinion.

I reverted it, since I specifically asked not to make this the
default.

Re: Failure building Emacs master

[Angelo Graziosi]

Il 21/03/2017 20:06, Eli Zaretskii ha scritto:
>> From: Paul Eggert <eggert@cs.ucla.edu>
>> Date: Tue, 21 Mar 2017 11:24:23 -0700
>> Cc: Emacs developers <emacs-devel@gnu.org>
>>
>> On 03/21/2017 09:45 AM, Angelo Graziosi wrote:
>>> It seems the patch is not yet applied on master..
>>
>> I applied it just now (with streamlined diagnostics), since you and I
>> liked the idea and there wasn't any opposing opinion.
>
> I reverted it, since I specifically asked not to make this the
> default.
>

..then you should fix also the configure warning... [*]


----
[*] http://lists.gnu.org/archive/html/emacs-devel/2017-03/msg00437.html

Re: Failure building Emacs master

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 515 bytes --]

>> I specifically asked not to make this the default.
>>

In <http://bugs.gnu.org/26102#39> you asked not to remove the POP source 
code. Sorry, I must have missed where you specifically asked not to make 
--without-pop the default (which is a different issue).

> ..then you should fix also the configure warning... [*] 

I installed the attached patch to try to improve the 'configure' 
diagnostics. Please feel free to revert or improve it if you think the 
patch makes this confusing area even more confusing.


[-- Attachment #2: 0001-Improve-configure-with-pop-etc.-diagnostics.patch --]
[-- Type: application/x-patch, Size: 2618 bytes --]

Re: Failure building Emacs master

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Tue, 21 Mar 2017 13:59:08 -0700
> 
> In <http://bugs.gnu.org/26102#39> you asked not to remove the POP source 
> code.

Not the POP source code, the fact that we build and install movemail
by default.

> I installed the attached patch to try to improve the 'configure' 
> diagnostics. Please feel free to revert or improve it if you think the 
> patch makes this confusing area even more confusing.

Thanks.  The message is fine with me, but I think Angelo should tell
if it's clear enough now.

Re: Failure building Emacs master

[Paul Eggert]

On 03/22/2017 09:10 AM, Eli Zaretskii wrote:
>> In<http://bugs.gnu.org/26102#39>  you asked not to remove the POP source
>> code.
> Not the POP source code, the fact that we build and install movemail
> by default.
>
Ah, that wasn't clear from your statement cited above. In 
<http://bugs.gnu.org/26102#30> I proposed one patch to remove the 
unencrypted POP3 source code from movemail, and another patch to add 
--with-movemail. Your statement vetoed the first and changed the second 
so that --with-movemail was the default. But neither patch made 
--without-pop the default; that was a third proposal I made later, a 
proposal that you did not respond to but merely reverted when it was 
eventually installed.

I'm writing all this because I don't want people to be left with the 
misimpression that I deliberately installed a patch that you had 
specifically asked not to be installed. That is not what happened.

Re: Failure building Emacs master

[Eli Zaretskii]

> Cc: angelo.graziosi@alice.it, emacs-devel@gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Wed, 22 Mar 2017 10:48:30 -0700
> 
> On 03/22/2017 09:10 AM, Eli Zaretskii wrote:
> >> In<http://bugs.gnu.org/26102#39>  you asked not to remove the POP source
> >> code.
> > Not the POP source code, the fact that we build and install movemail
> > by default.
> >
> Ah, that wasn't clear from your statement cited above.

Sorry about that, I thought it was clear.

> I'm writing all this because I don't want people to be left with the 
> misimpression that I deliberately installed a patch that you had 
> specifically asked not to be installed. That is not what happened.

Understood, thanks for clarifying.

Re: Failure building Emacs master

[Angelo Graziosi]

Il 22/03/2017 17:10, Eli Zaretskii ha scritto:
>> Cc: emacs-devel@gnu.org
>> From: Paul Eggert <eggert@cs.ucla.edu>
>> Date: Tue, 21 Mar 2017 13:59:08 -0700
>>
>> In <http://bugs.gnu.org/26102#39> you asked not to remove the POP source
>> code.
>
> Not the POP source code, the fact that we build and install movemail
> by default.
>
>> I installed the attached patch to try to improve the 'configure'
>> diagnostics. Please feel free to revert or improve it if you think the
>> patch makes this confusing area even more confusing.
>
> Thanks.  The message is fine with me, but I think Angelo should tell
> if it's clear enough now.
>

Really, I don't like the warning from 'configure' for a default feature. 
In any case now the message seem clear enough and you can "close" this 
"issue"..

Just out of curiosity, what would be installed if one uses "--without-pop"?

Thanks,
  Angelo.

Re: Failure building Emacs master

[Paul Eggert]

On 03/22/2017 11:59 AM, Angelo Graziosi wrote:

> I don't like the warning from 'configure' for a default feature.

Me neither. I'd rather change the default. However, as long as the 
default has a significant security issue we should at least warn the 
builder/installer about the issue.

> what would be installed if one uses "--without-pop"?

An Emacs-specific 'movemail' executable that copies mail from local 
mailboxes (presumably securely) but does not support (unencrypted-only) 
POP to retrieve email over a network.

Re: Failure building Emacs master

[Eli Zaretskii]

> From: Angelo Graziosi <angelo.graziosi@alice.it>
> Date: Wed, 22 Mar 2017 19:59:51 +0100
> Cc: emacs-devel@gnu.org
> 
> Really, I don't like the warning from 'configure' for a default feature. 

It's a first step towards removing the feature in some future
release.  How else would you suggest to warn users and distributions
about the issue without making drastic backward-incompatible changes
in the build products?

> In any case now the message seem clear enough and you can "close" this 
> "issue"..

Sorry, I don't understand the quotes.  If you are unhappy about the
solution, even after your questions were answered, let's try to find a
better one.

Re: Failure building Emacs master

[Angelo Graziosi]

Il 22/03/2017 19:59, Angelo Graziosi ha scritto:
>
> Just out of curiosity, what would be installed if one uses "--without-pop"?

I haven't yet understood what does "--without-pop"...

Is movemail built? How? what does it support?

> Thanks,
>  Angelo.

Re: Failure building Emacs master

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> From: Angelo Graziosi <angelo.graziosi@alice.it>
> Date: Sun, 26 Mar 2017 11:37:15 +0200
> 
> Il 22/03/2017 19:59, Angelo Graziosi ha scritto:
> >
> > Just out of curiosity, what would be installed if one uses "--without-pop"?
> 
> I haven't yet understood what does "--without-pop"...

Let me try to help.

> Is movemail built?

Yes.

> How?

As usual, by commands defined by lib-src/Makefile.

> what does it support?

It then supports only local mailboxes that are disk files.  As in
"var/mail/USER", for example.

Re: Failure building Emacs master

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> From: Angelo Graziosi <angelo.graziosi@alice.it>
> Date: Tue, 21 Mar 2017 21:13:37 +0100
> 
> Il 21/03/2017 20:06, Eli Zaretskii ha scritto:
> >> From: Paul Eggert <eggert@cs.ucla.edu>
> >> Date: Tue, 21 Mar 2017 11:24:23 -0700
> >> Cc: Emacs developers <emacs-devel@gnu.org>
> >>
> >> On 03/21/2017 09:45 AM, Angelo Graziosi wrote:
> >>> It seems the patch is not yet applied on master..
> >>
> >> I applied it just now (with streamlined diagnostics), since you and I
> >> liked the idea and there wasn't any opposing opinion.
> >
> > I reverted it, since I specifically asked not to make this the
> > default.
> >
> 
> ..then you should fix also the configure warning... [*]

I'm not against clarifying messages, I just think that removing a
feature which needs the message instead of improving the message is
too drastic a measure.

Paul clarified the message, so if the new wording is clear enough in
your opinion, we can close the issue.

Thanks.

Re: Failure building Emacs master

[Colin Baxter]

Hello,

On Wed, Mar 22 2017, Eli Zaretskii wrote:

>> Cc: emacs-devel@gnu.org
>> From: Angelo Graziosi <angelo.graziosi@alice.it>
>> Date: Tue, 21 Mar 2017 21:13:37 +0100
>> 
>> Il 21/03/2017 20:06, Eli Zaretskii ha scritto:
>> >> From: Paul Eggert <eggert@cs.ucla.edu>
>> >> Date: Tue, 21 Mar 2017 11:24:23 -0700
>> >> Cc: Emacs developers <emacs-devel@gnu.org>
>> >>
>> >> On 03/21/2017 09:45 AM, Angelo Graziosi wrote:
>> >>> It seems the patch is not yet applied on master..
>> >>
>> >> I applied it just now (with streamlined diagnostics), since you and I
>> >> liked the idea and there wasn't any opposing opinion.
>> >
>> > I reverted it, since I specifically asked not to make this the
>> > default.
>> >
>> 
>> ..then you should fix also the configure warning... [*]
>
> I'm not against clarifying messages, I just think that removing a
> feature which needs the message instead of improving the message is
> too drastic a measure.
>
> Paul clarified the message, so if the new wording is clear enough in
> your opinion, we can close the issue.
>
> Thanks.

I'm sorry to appear extremely stupid but I remain confused, even after
reading this thread again. If I use ./configure --with-mailutils then I
will (potentially) build emacs-26 with encrypted POP3 - correct?

Colin.

Re: Failure building Emacs master

[Eli Zaretskii]

> From: Colin Baxter <m43cap@yandex.com>
> Date: Wed, 22 Mar 2017 19:57:54 +0000
> Cc: , eggert@cs.ucla.edu, emacs-devel@gnu.org,
> 	Angelo Graziosi <angelo.graziosi@alice.it>
> 
> I'm sorry to appear extremely stupid but I remain confused, even after
> reading this thread again. If I use ./configure --with-mailutils then I
> will (potentially) build emacs-26 with encrypted POP3 - correct?

No, that option disables building movemail that comes with Emacs
altogether.  That option is for people who have GNU Mailutils
installed, which includes a much more powerful version of movemail,
including encrypted POP3 capability.

Re: Failure building Emacs master

[Colin Baxter]

On Wed, Mar 22 2017, Eli Zaretskii wrote:

>> From: Colin Baxter <m43cap@yandex.com>
>> Date: Wed, 22 Mar 2017 19:57:54 +0000
>> Cc: , eggert@cs.ucla.edu, emacs-devel@gnu.org,
>> 	Angelo Graziosi <angelo.graziosi@alice.it>
>> 
>> I'm sorry to appear extremely stupid but I remain confused, even after
>> reading this thread again. If I use ./configure --with-mailutils then I
>> will (potentially) build emacs-26 with encrypted POP3 - correct?
>
> No, that option disables building movemail that comes with Emacs
> altogether.  That option is for people who have GNU Mailutils
> installed, which includes a much more powerful version of movemail,
> including encrypted POP3 capability.

Thank you. I see where I was confused.  I have GNU Mailutils installed
so I will therefore use ./configure --with-mailutils to obtain POP3
encryption.

Thanks again,

Colin.

Re: Failure building Emacs master

[Angelo Graziosi]

Paul Eggert wrote:
> If so, the attached patch should implement your suggestion.
>
> PS. I fixed the glitch with ../nt that you noted.

I have built with that patch and all seems OK: no failure nor warning..

Thanks,
  Angelo.

Failure building Emacs master

[Angelo Graziosi]

Trying to build master r128650.3139914f51 on MSYS2/MinGW64 fails as follows:

MINGW_INSTALLS=mingw64 makepkg-mingw -sLf
[...]
==> Avvio di prepare() in corso...
patching file src/image.c
Hunk #1 succeeded at 8206 (offset 344 lines).
patching file src/lread.c
Hunk #1 succeeded at 4418 (offset 67 lines).
Checking whether you have the necessary tools...
(Read INSTALL.REPO for more details on building Emacs)
Checking for autoconf (need at least version 2.65) ... ok
Your system has the required tools.
Building aclocal.m4 ...
Running 'autoreconf -fi -I m4' ...
You can now run './autogen.sh git'.
==> Avvio di pkgver() in corso...
==> Versione aggiornata: mingw-w64-emacs-git r128650.3139914f51-1
==> Avvio di build() in corso...
configure: loading site script /etc/config.site
checking for xcrun... no
checking for GNU Make... make
checking build system type... x86_64-w64-mingw32
checking host system type... x86_64-w64-mingw32
checking the compiler's target... x86_64-w64-mingw32
[...]
Configured for 'x86_64-w64-mingw32'.

   Where should the build process find the source code? 
/C/msys64/tmp/mingw-w64-emacs-git/src/emacs
   What compiler should emacs be built with?               gcc 
-march=x86-64 -mtune=generic -O2 -pipe
   Should Emacs use the GNU version of malloc?             no
     (The GNU allocators don't work with this system configuration.)
   Should Emacs use a relocating allocator for buffers?    no
   Should Emacs use mmap(2) for buffer allocation?         yes
   What window system should Emacs use?                    w32
[...]
config.status: executing etc-refcards-emacsver.tex commands
configure: WARNING: This configuration installs a 'movemail' program
that retrieves POP3 email via only insecure channels.
To fix this you can use '/tmp/mingw-w64-emacs-git/src/emacs/configure 
--without-pop'.

make -C nt all
make -C lib all
make info-real info-dir
make[1]: ingresso nella directory 
"/tmp/mingw-w64-emacs-git/src/build-x86_64-w64-mingw32/nt"
make[1]: ingresso nella directory 
"/tmp/mingw-w64-emacs-git/src/build-x86_64-w64-mingw32/lib"
Makefile:60: ../nt/gnulib-cfg.mk: No such file or directory
make[1]: ***  Nessuna regola per generare l'obiettivo 
"../nt/gnulib-cfg.mk".  Arresto.
make[1]: uscita dalla directory 
"/tmp/mingw-w64-emacs-git/src/build-x86_64-w64-mingw32/lib"
make: *** [Makefile:403: lib] Error 2
make: *** Attesa per i processi non terminati....
   RC       emacs.res
   CCLD     addpm.exe
make[1]: ingresso nella directory 
"/tmp/mingw-w64-emacs-git/src/build-x86_64-w64-mingw32"
make -C doc/lispref info
make[2]: ingresso nella directory 
"/tmp/mingw-w64-emacs-git/src/build-x86_64-w64-mingw32/doc/lispref"
/usr/bin/mkdir -p 
/C/msys64/tmp/mingw-w64-emacs-git/src/emacs/doc/lispref/../../info
   GEN 
/C/msys64/tmp/mingw-w64-emacs-git/src/emacs/doc/lispref/../../info/elisp.info
   CCLD     cmdproxy.exe
[...]

Beside this, I don't understand this warning:

configure: WARNING: This configuration installs a 'movemail' program
that retrieves POP3 email via only insecure channels.
To fix this you can use '/tmp/mingw-w64-emacs-git/src/emacs/configure 
--without-pop'.

You should have a configure with enable by default. If one wants POP 
then it will configure --with-pop..

Angelo

Re: Failure building Emacs master

[Angelo Graziosi]

Il 18/03/2017 00:03, Angelo Graziosi ha scritto:

>
> Beside this, I don't understand this warning:
>
> configure: WARNING: This configuration installs a 'movemail' program
> that retrieves POP3 email via only insecure channels.
> To fix this you can use '/tmp/mingw-w64-emacs-git/src/emacs/configure
> --without-pop'.
>

Oops...

> You should have a configure with enable by default. If one wants POP
> then it will configure --with-pop..

Maybe this is better:

You should have a configure with '--without-pop' enabled by default. If 
one wants POP then it will configure --with-pop..

>
> Angelo
>

Re: Failure building Emacs master

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 484 bytes --]

On 03/17/2017 04:07 PM, Angelo Graziosi wrote:
>
> You should have a configure with '--without-pop' enabled by default. 
> If one wants POP then it will configure --with-pop.. 

Although I originally implemented it that way, Eli worried that it was 
too bug of a change. Perhaps he will reconsider, given that --with-pop 
is evidently confusing as well as being insecure. If so, the attached 
patch should implement your suggestion.

PS. I fixed the glitch with ../nt that you noted.

[-- Attachment #2: 0001-Make-without-pop-the-default.patch --]
[-- Type: text/x-patch, Size: 4150 bytes --]

From 9baa037d5afb111676a012a64fb8476853f43ddf Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Fri, 17 Mar 2017 21:57:09 -0700
Subject: [PATCH] Make --without-pop the default.

Suggested by Angelo Graziosi in:
http://lists.gnu.org/archive/html/emacs-devel/2017-03/msg00431.html
* configure.ac: Change the default from --with-pop to
--without-pop.
---
 INSTALL      |  7 ++-----
 configure.ac | 11 ++++-------
 etc/NEWS     | 13 ++++++++++---
 3 files changed, 16 insertions(+), 15 deletions(-)

diff --git a/INSTALL b/INSTALL
index a99f107..9b6c04f 100644
--- a/INSTALL
+++ b/INSTALL
@@ -265,17 +265,14 @@ uses the GNU Mailutils 'movemail' program to retrieve mail.  Otherwise
 the Emacs build procedure builds and installs an auxiliary 'movemail'
 program, a limited and insecure substitute that Emacs can use when
 Mailutils is not installed; when this happens, there are several
-configure options such as --without-pop that provide fine-grained
+configure options such as --with-mail-unlink that provide fine-grained
 control over Emacs 'movemail' construction.
 
 The Emacs mail reader RMAIL is configured to be able to read mail from
 a POP3 server by default.  Versions of the POP protocol older than
 POP3 are not supported.  While POP3 support is typically enabled,
 whether Emacs actually uses POP3 is controlled by individual users;
-see the Rmail chapter of the Emacs manual.  Unless you configure
---with-mailutils, it is a good idea to configure --without-pop so that
-users are less likely to inadvertently read email via insecure
-channels.
+see the Rmail chapter of the Emacs manual.
 
 For image support you may have to download, build, and install the
 appropriate image support libraries for image types other than XBM and
diff --git a/configure.ac b/configure.ac
index b922dc4..2712333 100644
--- a/configure.ac
+++ b/configure.ac
@@ -229,19 +229,16 @@ AC_DEFUN
    m4_bpatsubst([with_$1], [[^0-9a-z]], [_])=$with_features])dnl
 ])dnl
 
-# FIXME: The default options '--without-mailutils --with-pop' result
-# in a movemail implementation that supports only unencrypted POP3
-# connections.  Encrypted connections should be the default.
-
 OPTION_DEFAULT_OFF([mailutils],
   [rely on GNU Mailutils being installed; this makes the options
-   --without-pop through --with-mailhost irrelevant])
+   --with-pop through --with-mailhost irrelevant])
 if test "$with_mailutils" = no; then
   with_mailutils=
 fi
 AC_SUBST([with_mailutils])
 
-OPTION_DEFAULT_ON([pop],[don't support POP mail retrieval with movemail])
+OPTION_DEFAULT_OFF([pop],
+  [support unencrypted-only POP mail retrieval with movemail])
 if test "$with_pop" = yes; then
    AC_DEFINE(MAIL_USE_POP)
 fi
@@ -5477,7 +5474,7 @@ m4_define
     cygwin | mingw32)
       # Don't suggest GNU Mailutils, as it hasn't been ported.
       if test "$with_pop" = yes; then
-	emacs_fix_movemail="use '$0 --without-pop'"
+	emacs_fix_movemail="use '$0' (without '--with-pop')"
       else
 	emacs_fix_movemail=
       fi;;
diff --git a/etc/NEWS b/etc/NEWS
index f047aa8..1548824 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -39,9 +39,16 @@ version 2.6.6 or later.
 
 ** The new option 'configure --with-mailutils' causes Emacs to rely on
 GNU Mailutils 'movemail' to retrieve email.  By default, the Emacs
-build procedure continues to build and install a limited and insecure
-'movemail' substitute.  Although --with-mailutils is recommended, it
-is not yet the default due to backward-compatibility concerns.
+build procedure continues to build and install a limited and
+possibly-insecure 'movemail' substitute.  Although --with-mailutils is
+recommended, it is not yet the default due to backward-compatibility
+concerns.
+
+** The option 'configure --without-pop' is now the default.
+(This option's setting matters only if --with-mailutils is not used.)
+Using --with-pop (the previous default) preserves the historical
+behavior where Emacs 'movemail' retrieves network email only via
+insecure channels.
 
 ** The new option 'configure --enable-gcc-warnings=warn-only' causes
 GCC to issue warnings without stopping the build.  This behavior is
-- 
2.9.3