Re: [pretest] emacs-26.0.90

[Grégoire Jadi]

Piotr Isajew  writes:

> On Mon, Nov 06, 2017 at 03:24:38PM +0100, Grégoire Jadi wrote:
>
>> Jeremie Courreges-Anglas  writes:
>> 
>> > Very interesting, thanks to both of you for your reports and
>> > investigations.  I've never hit this, probably because I run the no_x11
>> > flavor.  I'll see what I can do to get this fixed, but if someone (tm)
>> > affected can file a bugreport upstream, that would be great. :)
>> 
>> Done, I've reported this bug upstream.
>> 
>> Piotr, if you think this is the same bug, you can add your backtraces to
>> it by sending a mail to 29...@debbugs.gnu.org.
>
> I've added my input to your report Grégoire. 

Thank you.

>> 
>> Best,

Re: [pretest] emacs-26.0.90

[Grégoire Jadi]

Jeremie Courreges-Anglas  writes:

> Very interesting, thanks to both of you for your reports and
> investigations.  I've never hit this, probably because I run the no_x11
> flavor.  I'll see what I can do to get this fixed, but if someone (tm)
> affected can file a bugreport upstream, that would be great. :)

Done, I've reported this bug upstream.

Piotr, if you think this is the same bug, you can add your backtraces to
it by sending a mail to 29...@debbugs.gnu.org.

Best,

Re: [pretest] emacs-26.0.90

[Jeremie Courreges-Anglas]

On Mon, Nov 06 2017, gj...@omecha.info (Grégoire Jadi) wrote:
> Piotr Isajew  writes:
>
>> Dnia 15.10.2017 Jeremie Courreges-Anglas  napisał/a:
>>>
>>> Here's a diff for the first pretest in the emacs-26 release cycle.
>>> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
>>> is still used at build/unexec time).
>>>
>>> As usual build & test reports are welcome, especially if you use an arch
>>> different from amd64 / arm, or if you use one of the graphical flavors.
>>
>> I've been testing this on 6.2. I notice one problem which exists
>> also for emacs-25 and I can reproduce it for all the graphical
>> flavors.
>>
>> The scenario I can reproduce is related to using org-capture. On
>> any other platform, or even on OpenBSD without X, after selecting
>> capture template a dedicated buffer appears in which I can edit
>> the template. On OpenBSD however, under X this buffer does only
>> appear after generating additional input event, like moving a
>> mouse. There's no such problem if I select capture template using
>> soft-keyboard - XVkbd.
>>
>> Also it is very likely that after selecting the capture template
>> *emacs will hang in kind of an infinite loop* maxing out CPU
>> usage.
>
> Hello,

Hi Piotr, salut Grégoire,

> I think I have encountered the same issue. The problem occurs when emacs
> tries to read the secondary clipbard with the encoding iso-8859-1.
> You can try to reproduce the bug with :
>> (setf x-selection-timeout 500)
>> (org-get-x-clipboard 'SECONDARY)
>
> If this is the same bug, the call to `org-get-x-clipboard' will froze
> emacs for 500ms then return nil.
>
> After I track this down I found a quick-fix :
>> (setf x-selection-timeout 10)
> This set the timeout to 10ms instead of 0 (i.e. wait indefinitely)
>
> At the time, I ran out of patience and didn't bother to find out if the
> bug was specific to OpenBSD or some weird interactions between emacs and
> my window manager and/or X. But if you confirm the same bug it may be
> time to report it.

Very interesting, thanks to both of you for your reports and
investigations.  I've never hit this, probably because I run the no_x11
flavor.  I'll see what I can do to get this fixed, but if someone (tm)
affected can file a bugreport upstream, that would be great. :)

> Best,
>
> PS: You can usually recover from a frozen emacs by sending the SIGUSR2
> signal to the emacs' process (pkill -USR2 emacs). When it receives
> SIGUSR2, emacs will drop into its debugger and toggle debug-on-quit
> (SIGUSR2 is like C-g but emacs should react to it even if the input loop
> is frozen).
>
>> The command I used to build the port:
>>
>> cd /usr/ports/editors/emacs
>> doas sh -c 'export FLAVOR=athena; export DEBUG="-g -O0 -DTRACE_SELECTION";
>>  make clean repackage reinstall'
>>
>> The stacktrace from the moment it hangs:
>>
>> info thre
>>   4 thread 424274  _thread_sys_poll () at -:3
>>   3 thread 437321  _thread_sys_poll () at -:3
>> * 2 thread 476142  _thread_sys_poll () at -:3
>>   1 thread 131181  wait_reading_process_output (time_limit=5, nsecs=0, 
>> read_kbd=0, do_display=false, wait_for_cell=16374659, wait_proc=0x0, 
>> just_wait_proc=0) at process.c:5296
>> (gdb) thr 4
>> [Switching to thread 4 (thread 424274)]#0  _thread_sys_poll () at -:3
>> 3   in -
>> (gdb) inf sta
>> #0  _thread_sys_poll () at -:3
>> #1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
>> available.
>> )
>> at /usr/src/lib/libc/sys/w_poll.c:27
>> #2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
>> #3  0x00022d883594 in g_main_context_iteration (context=0x20ddb0500, 
>> may_block=The value of variable 'may_block' is distributed across several
>> locations, and GDB cannot access its value.
>>
>> ) at gmain.c:4033
>> #4  0x0002c7f2bb7d in g_io_module_query ()
>>from /usr/local/lib/gio/modules/libdconfsettings.so
>> #5  0x00022d8ac3fa in g_thread_proxy (data=0x259d50190) at gthread.c:784
>> #6  0x000275642cae in _rthread_start (v=Variable "v" is not available.
>> )
>> at /usr/src/lib/librthread/rthread.c:96
>> #7  0x0002c6d1b96b in __tfork_thread ()
>> at /usr/src/lib/libc/arch/amd64/sys/tfork_thread.S:75
>> #8  0x in ?? ()
>>
>> (gdb) thr 3
>> [Switching to thread 3 (thread 437321)]#0  _thread_sys_poll () at -:3
>> 3   in -
>> (gdb) inf sta
>> #0  _thread_sys_poll () at -:3
>> #1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
>> available.
>> )
>> at /usr/src/lib/libc/sys/w_poll.c:27
>> #2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
>> #3  0x00022d883594 in g_main_context_iteration (context=0x243a3ce00, 
>> may_block=The value of variable 'may_block' is distributed across several
>> locations, and GDB cannot access its value.
>>
>> ) at gmain.c:4033
>> #4  0x00022d885296 in glib_worker_main (data=Variable "data" is not 
>> available.
>> ) at gmain.c:5824
>> 

Re: [pretest] emacs-26.0.90

[Grégoire Jadi]

Piotr Isajew  writes:

> Dnia 15.10.2017 Jeremie Courreges-Anglas  napisał/a:
>>
>> Here's a diff for the first pretest in the emacs-26 release cycle.
>> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
>> is still used at build/unexec time).
>>
>> As usual build & test reports are welcome, especially if you use an arch
>> different from amd64 / arm, or if you use one of the graphical flavors.
>
> I've been testing this on 6.2. I notice one problem which exists
> also for emacs-25 and I can reproduce it for all the graphical
> flavors.
>
> The scenario I can reproduce is related to using org-capture. On
> any other platform, or even on OpenBSD without X, after selecting
> capture template a dedicated buffer appears in which I can edit
> the template. On OpenBSD however, under X this buffer does only
> appear after generating additional input event, like moving a
> mouse. There's no such problem if I select capture template using
> soft-keyboard - XVkbd.
>
> Also it is very likely that after selecting the capture template
> *emacs will hang in kind of an infinite loop* maxing out CPU
> usage.

Hello,

I think I have encountered the same issue. The problem occurs when emacs
tries to read the secondary clipbard with the encoding iso-8859-1.
You can try to reproduce the bug with :
> (setf x-selection-timeout 500)
> (org-get-x-clipboard 'SECONDARY)

If this is the same bug, the call to `org-get-x-clipboard' will froze
emacs for 500ms then return nil.

After I track this down I found a quick-fix :
> (setf x-selection-timeout 10)
This set the timeout to 10ms instead of 0 (i.e. wait indefinitely)

At the time, I ran out of patience and didn't bother to find out if the
bug was specific to OpenBSD or some weird interactions between emacs and
my window manager and/or X. But if you confirm the same bug it may be
time to report it.

Best,

PS: You can usually recover from a frozen emacs by sending the SIGUSR2
signal to the emacs' process (pkill -USR2 emacs). When it receives
SIGUSR2, emacs will drop into its debugger and toggle debug-on-quit
(SIGUSR2 is like C-g but emacs should react to it even if the input loop
is frozen).

> The command I used to build the port:
>
> cd /usr/ports/editors/emacs
> doas sh -c 'export FLAVOR=athena; export DEBUG="-g -O0 -DTRACE_SELECTION";
>  make clean repackage reinstall'
>
> The stacktrace from the moment it hangs:
>
> info thre
>   4 thread 424274  _thread_sys_poll () at -:3
>   3 thread 437321  _thread_sys_poll () at -:3
> * 2 thread 476142  _thread_sys_poll () at -:3
>   1 thread 131181  wait_reading_process_output (time_limit=5, nsecs=0, 
> read_kbd=0, do_display=false, wait_for_cell=16374659, wait_proc=0x0, 
> just_wait_proc=0) at process.c:5296
> (gdb) thr 4
> [Switching to thread 4 (thread 424274)]#0  _thread_sys_poll () at -:3
> 3   in -
> (gdb) inf sta
> #0  _thread_sys_poll () at -:3
> #1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
> available.
> )
> at /usr/src/lib/libc/sys/w_poll.c:27
> #2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
> #3  0x00022d883594 in g_main_context_iteration (context=0x20ddb0500, 
> may_block=The value of variable 'may_block' is distributed across several
> locations, and GDB cannot access its value.
>
> ) at gmain.c:4033
> #4  0x0002c7f2bb7d in g_io_module_query ()
>from /usr/local/lib/gio/modules/libdconfsettings.so
> #5  0x00022d8ac3fa in g_thread_proxy (data=0x259d50190) at gthread.c:784
> #6  0x000275642cae in _rthread_start (v=Variable "v" is not available.
> )
> at /usr/src/lib/librthread/rthread.c:96
> #7  0x0002c6d1b96b in __tfork_thread ()
> at /usr/src/lib/libc/arch/amd64/sys/tfork_thread.S:75
> #8  0x in ?? ()
>
> (gdb) thr 3
> [Switching to thread 3 (thread 437321)]#0  _thread_sys_poll () at -:3
> 3   in -
> (gdb) inf sta
> #0  _thread_sys_poll () at -:3
> #1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
> available.
> )
> at /usr/src/lib/libc/sys/w_poll.c:27
> #2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
> #3  0x00022d883594 in g_main_context_iteration (context=0x243a3ce00, 
> may_block=The value of variable 'may_block' is distributed across several
> locations, and GDB cannot access its value.
>
> ) at gmain.c:4033
> #4  0x00022d885296 in glib_worker_main (data=Variable "data" is not 
> available.
> ) at gmain.c:5824
> #5  0x00022d8ac3fa in g_thread_proxy (data=0x259d5) at gthread.c:784
> #6  0x000275642cae in _rthread_start (v=Variable "v" is not available.
> )
> at /usr/src/lib/librthread/rthread.c:96
> #7  0x0002c6d1b96b in __tfork_thread ()
> at /usr/src/lib/libc/arch/amd64/sys/tfork_thread.S:75
>
> (gdb) thr 2
> [Switching to thread 2 (thread 476142)]#0  _thread_sys_poll () at -:3
> 3   in -
> (gdb) inf sta
> #0  _thread_sys_poll () at -:3
> #1  0x0002c6d1f45f 

Re: [pretest] emacs-26.0.90

[Piotr Isajew]

Dnia 15.10.2017 Jeremie Courreges-Anglas  napisał/a:
>
> Here's a diff for the first pretest in the emacs-26 release cycle.
> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
> is still used at build/unexec time).
>
> As usual build & test reports are welcome, especially if you use an arch
> different from amd64 / arm, or if you use one of the graphical flavors.

I've been testing this on 6.2. I notice one problem which exists
also for emacs-25 and I can reproduce it for all the graphical
flavors.

The scenario I can reproduce is related to using org-capture. On
any other platform, or even on OpenBSD without X, after selecting
capture template a dedicated buffer appears in which I can edit
the template. On OpenBSD however, under X this buffer does only
appear after generating additional input event, like moving a
mouse. There's no such problem if I select capture template using
soft-keyboard - XVkbd.

Also it is very likely that after selecting the capture template
*emacs will hang in kind of an infinite loop* maxing out CPU
usage.


The command I used to build the port:

cd /usr/ports/editors/emacs
doas sh -c 'export FLAVOR=athena; export DEBUG="-g -O0 -DTRACE_SELECTION";
 make clean repackage reinstall'


The stacktrace from the moment it hangs:

info thre
  4 thread 424274  _thread_sys_poll () at -:3
  3 thread 437321  _thread_sys_poll () at -:3
* 2 thread 476142  _thread_sys_poll () at -:3
  1 thread 131181  wait_reading_process_output (time_limit=5, nsecs=0, 
read_kbd=0, do_display=false, wait_for_cell=16374659, wait_proc=0x0, 
just_wait_proc=0) at process.c:5296
(gdb) thr 4
[Switching to thread 4 (thread 424274)]#0  _thread_sys_poll () at -:3
3   in -
(gdb) inf sta
#0  _thread_sys_poll () at -:3
#1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
available.
)
at /usr/src/lib/libc/sys/w_poll.c:27
#2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
#3  0x00022d883594 in g_main_context_iteration (context=0x20ddb0500, 
may_block=The value of variable 'may_block' is distributed across several
locations, and GDB cannot access its value.

) at gmain.c:4033
#4  0x0002c7f2bb7d in g_io_module_query ()
   from /usr/local/lib/gio/modules/libdconfsettings.so
#5  0x00022d8ac3fa in g_thread_proxy (data=0x259d50190) at gthread.c:784
#6  0x000275642cae in _rthread_start (v=Variable "v" is not available.
)
at /usr/src/lib/librthread/rthread.c:96
#7  0x0002c6d1b96b in __tfork_thread ()
at /usr/src/lib/libc/arch/amd64/sys/tfork_thread.S:75
#8  0x in ?? ()

(gdb) thr 3
[Switching to thread 3 (thread 437321)]#0  _thread_sys_poll () at -:3
3   in -
(gdb) inf sta
#0  _thread_sys_poll () at -:3
#1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
available.
)
at /usr/src/lib/libc/sys/w_poll.c:27
#2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
#3  0x00022d883594 in g_main_context_iteration (context=0x243a3ce00, 
may_block=The value of variable 'may_block' is distributed across several
locations, and GDB cannot access its value.

) at gmain.c:4033
#4  0x00022d885296 in glib_worker_main (data=Variable "data" is not 
available.
) at gmain.c:5824
#5  0x00022d8ac3fa in g_thread_proxy (data=0x259d5) at gthread.c:784
#6  0x000275642cae in _rthread_start (v=Variable "v" is not available.
)
at /usr/src/lib/librthread/rthread.c:96
#7  0x0002c6d1b96b in __tfork_thread ()
at /usr/src/lib/libc/arch/amd64/sys/tfork_thread.S:75

(gdb) thr 2
[Switching to thread 2 (thread 476142)]#0  _thread_sys_poll () at -:3
3   in -
(gdb) inf sta
#0  _thread_sys_poll () at -:3
#1  0x0002c6d1f45f in _libc_poll_cancel (fds=Variable "fds" is not 
available.
)
at /usr/src/lib/libc/sys/w_poll.c:27
#2  0x00022d883497 in g_main_context_iterate () at gmain.c:4271
#3  0x00022d88382f in g_main_loop_run (loop=0x249a70a80) at gmain.c:4168
#4  0x0002a5dac8cb in gdbus_shared_thread_func (user_data=0x2039baa80)
at gdbusprivate.c:252
#5  0x00022d8ac3fa in g_thread_proxy (data=0x27f39fed0) at gthread.c:784
#6  0x000275642cae in _rthread_start (v=Variable "v" is not available.
)
at /usr/src/lib/librthread/rthread.c:96
#7  0x0002c6d1b96b in __tfork_thread ()
at /usr/src/lib/libc/arch/amd64/sys/tfork_thread.S:75
#8  0x in ?? ()


(gdb) thr 1
[Switching to thread 1 (thread 131181)]#0  wait_reading_process_output (
time_limit=5, nsecs=0, read_kbd=0, do_display=false, 
wait_for_cell=16374659, wait_proc=0x0, just_wait_proc=0) at process.c:5296
5296  FD_ZERO ();
Current language:  auto; currently minimal
(gdb) inf sta
#0  wait_reading_process_output (time_limit=5, nsecs=0, read_kbd=0, 
do_display=false, wait_for_cell=16374659, wait_proc=0x0, just_wait_proc=0)
at process.c:5296
#1  0x005b297d in x_get_foreign_selection (selection_symbol=9120, 

Re: [pretest] emacs-26.0.90

[Grégoire Jadi]

Jeremie Courreges-Anglas  writes:

> On Thu, Oct 26 2017, Jeremie Courreges-Anglas  wrote:
>> On Wed, Oct 25 2017, Jeremie Courreges-Anglas  wrote:
>>> On Sun, Oct 15 2017, Jeremie Courreges-Anglas  wrote:
 Here's a diff for the first pretest in the emacs-26 release cycle.
 Among other things, emacs now uses system malloc(3) at runtime (gmalloc
 is still used at build/unexec time).

 As usual build & test reports are welcome, especially if you use an arch
 different from amd64 / arm, or if you use one of the graphical flavors.
>>>
>>> Thanks for the feedback so far, updated diff (sync with -current, use
>>> https instead of ftp).  I think we're going to stay with .xz tarballs
>>> for next releases, ~40M vs ~60M.
>>>
>>> The build is broken on armv7.  Tests on any arch you have at hand are
>>> still welcome.
>>
>> The previous patch was messed up, sorry about that.
>
> Fixed the gtk3 flavor, with an additional patch from upstream to unbreak
> on arm.
Hello,

gtk3 on amd64 builds and runs correctly.

Thanks.

> Index: Makefile
> ===
> RCS file: /d/cvs/ports/editors/emacs/Makefile,v
> retrieving revision 1.71
> diff -u -p -r1.71 Makefile
> --- Makefile  19 Oct 2017 16:28:48 -  1.71
> +++ Makefile  27 Oct 2017 15:44:36 -
> @@ -2,8 +2,7 @@
>
>  COMMENT= GNU editor: extensible, customizable, self-documenting
>
> -VERSION= 25.3
> -REVISION=0
> +VERSION= 26.0.90
>  DISTNAME=emacs-${VERSION}
>
>  CATEGORIES=  editors
> @@ -17,7 +16,8 @@ PERMIT_PACKAGE_CDROM=   Yes
>
>  WANTLIB= c m ncurses pthread gnutls xml2 z
>
> -MASTER_SITES=${MASTER_SITE_GNU:=emacs/}
> +MASTER_SITES=https://alpha.gnu.org/gnu/emacs/pretest/
> +EXTRACT_SUFX=.tar.xz
>
>  USE_GMAKE=   Yes
>
> @@ -47,46 +47,47 @@ LIB_DEPENDS=  security/gnutls \
>  .if ${FLAVOR} == "no_x11"
>  CONFIGURE_ARGS+= --without-x \
>   --without-dbus \
> - --without-gconf \
>   --without-gsettings \
> - --without-jpeg
> + --without-jpeg \
> + --without-lcms2
>  .else
>  LIB_DEPENDS+=x11/dbus \
>   x11/gnome/librsvg \
> - devel/gconf2 \
>   graphics/jpeg \
>   graphics/png \
>   graphics/tiff \
>   graphics/giflib \
> - graphics/ImageMagick
> + graphics/ImageMagick \
> + graphics/lcms2
>  RUN_DEPENDS+=devel/desktop-file-utils \
>   devel/xdg-utils \
>   x11/gtk+3,-guic
>  . if ${FLAVOR} == "athena"
>  CONFIGURE_ARGS+= --with-x-toolkit=athena
>  LIB_DEPENDS+=x11/Xaw3d
> -WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 Xaw3d
> -WANTLIB += Xext Xft Xinerama Xmu Xpm Xrandr Xrender Xt cairo dbus-1
> -WANTLIB += fontconfig freetype gconf-2 gdk_pixbuf-2.0 gif gio-2.0
> -WANTLIB += glib-2.0 gobject-2.0 intl jpeg png rsvg-2 tiff X11-xcb Xfixes xcb
> +WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
> +WANTLIB += Xaw3d Xext Xfixes Xft Xinerama Xmu Xpm Xrandr Xrender
> +WANTLIB += Xt cairo dbus-1 fontconfig freetype gdk_pixbuf-2.0
> +WANTLIB += gif gio-2.0 glib-2.0 gobject-2.0 intl jpeg lcms2 png
> +WANTLIB += rsvg-2 tiff xcb
>  . elif ${FLAVOR} == "gtk2"
>  CONFIGURE_ARGS+= --with-x-toolkit=gtk2
>  LIB_DEPENDS+=x11/gtk+2
> -WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 Xcomposite
> -WANTLIB += Xcursor Xdamage Xext Xfixes Xft Xi Xinerama Xpm Xrandr
> -WANTLIB += Xrender atk-1.0 cairo dbus-1 fontconfig freetype gconf-2
> -WANTLIB += gdk-x11-2.0 gdk_pixbuf-2.0 gif gio-2.0 glib-2.0 gobject-2.0
> -WANTLIB += gtk-x11-2.0 intl jpeg pango-1.0 pangocairo-1.0 pangoft2-1.0
> -WANTLIB += png rsvg-2 tiff X11-xcb xcb
> +WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
> +WANTLIB += Xcomposite Xcursor Xdamage Xext Xfixes Xft Xi Xinerama
> +WANTLIB += Xpm Xrandr Xrender atk-1.0 cairo dbus-1 fontconfig
> +WANTLIB += freetype gdk-x11-2.0 gdk_pixbuf-2.0 gif gio-2.0 glib-2.0
> +WANTLIB += gobject-2.0 gtk-x11-2.0 intl jpeg lcms2 pango-1.0 pangocairo-1.0
> +WANTLIB += pangoft2-1.0 png rsvg-2 tiff xcb
>  . elif ${FLAVOR} == "gtk3"
>  CONFIGURE_ARGS+= --with-x-toolkit=gtk3
>  LIB_DEPENDS+=x11/gtk+3
>  WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
> -WANTLIB += Xfixes Xft Xinerama Xpm Xrandr Xrender atk-1.0 cairo
> -WANTLIB += cairo-gobject dbus-1 fontconfig freetype gconf-2 gdk-3
> +WANTLIB += Xext Xfixes Xft Xinerama Xpm Xrandr Xrender atk-1.0
> +WANTLIB += cairo cairo-gobject dbus-1 fontconfig 

Re: [pretest] emacs-26.0.90

[Jeremie Courreges-Anglas]

On Thu, Oct 26 2017, Jeremie Courreges-Anglas  wrote:
> On Wed, Oct 25 2017, Jeremie Courreges-Anglas  wrote:
>> On Sun, Oct 15 2017, Jeremie Courreges-Anglas  wrote:
>>> Here's a diff for the first pretest in the emacs-26 release cycle.
>>> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
>>> is still used at build/unexec time).
>>>
>>> As usual build & test reports are welcome, especially if you use an arch
>>> different from amd64 / arm, or if you use one of the graphical flavors.
>>
>> Thanks for the feedback so far, updated diff (sync with -current, use
>> https instead of ftp).  I think we're going to stay with .xz tarballs
>> for next releases, ~40M vs ~60M.
>>
>> The build is broken on armv7.  Tests on any arch you have at hand are
>> still welcome.
>
> The previous patch was messed up, sorry about that.

Fixed the gtk3 flavor, with an additional patch from upstream to unbreak
on arm.


Index: Makefile
===
RCS file: /d/cvs/ports/editors/emacs/Makefile,v
retrieving revision 1.71
diff -u -p -r1.71 Makefile
--- Makefile19 Oct 2017 16:28:48 -  1.71
+++ Makefile27 Oct 2017 15:44:36 -
@@ -2,8 +2,7 @@
 
 COMMENT=   GNU editor: extensible, customizable, self-documenting
 
-VERSION=   25.3
-REVISION=  0
+VERSION=   26.0.90
 DISTNAME=  emacs-${VERSION}
 
 CATEGORIES=editors
@@ -17,7 +16,8 @@ PERMIT_PACKAGE_CDROM= Yes
 
 WANTLIB=   c m ncurses pthread gnutls xml2 z
 
-MASTER_SITES=  ${MASTER_SITE_GNU:=emacs/}
+MASTER_SITES=  https://alpha.gnu.org/gnu/emacs/pretest/
+EXTRACT_SUFX=  .tar.xz
 
 USE_GMAKE= Yes
 
@@ -47,46 +47,47 @@ LIB_DEPENDS=security/gnutls \
 .if ${FLAVOR} == "no_x11"
 CONFIGURE_ARGS+=   --without-x \
--without-dbus \
-   --without-gconf \
--without-gsettings \
-   --without-jpeg
+   --without-jpeg \
+   --without-lcms2
 .else
 LIB_DEPENDS+=  x11/dbus \
x11/gnome/librsvg \
-   devel/gconf2 \
graphics/jpeg \
graphics/png \
graphics/tiff \
graphics/giflib \
-   graphics/ImageMagick
+   graphics/ImageMagick \
+   graphics/lcms2
 RUN_DEPENDS+=  devel/desktop-file-utils \
devel/xdg-utils \
x11/gtk+3,-guic
 . if ${FLAVOR} == "athena"
 CONFIGURE_ARGS+=   --with-x-toolkit=athena
 LIB_DEPENDS+=  x11/Xaw3d
-WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 Xaw3d
-WANTLIB += Xext Xft Xinerama Xmu Xpm Xrandr Xrender Xt cairo dbus-1
-WANTLIB += fontconfig freetype gconf-2 gdk_pixbuf-2.0 gif gio-2.0
-WANTLIB += glib-2.0 gobject-2.0 intl jpeg png rsvg-2 tiff X11-xcb Xfixes xcb
+WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
+WANTLIB += Xaw3d Xext Xfixes Xft Xinerama Xmu Xpm Xrandr Xrender
+WANTLIB += Xt cairo dbus-1 fontconfig freetype gdk_pixbuf-2.0
+WANTLIB += gif gio-2.0 glib-2.0 gobject-2.0 intl jpeg lcms2 png
+WANTLIB += rsvg-2 tiff xcb
 . elif ${FLAVOR} == "gtk2"
 CONFIGURE_ARGS+=   --with-x-toolkit=gtk2
 LIB_DEPENDS+=  x11/gtk+2
-WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 Xcomposite
-WANTLIB += Xcursor Xdamage Xext Xfixes Xft Xi Xinerama Xpm Xrandr
-WANTLIB += Xrender atk-1.0 cairo dbus-1 fontconfig freetype gconf-2
-WANTLIB += gdk-x11-2.0 gdk_pixbuf-2.0 gif gio-2.0 glib-2.0 gobject-2.0
-WANTLIB += gtk-x11-2.0 intl jpeg pango-1.0 pangocairo-1.0 pangoft2-1.0
-WANTLIB += png rsvg-2 tiff X11-xcb xcb
+WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
+WANTLIB += Xcomposite Xcursor Xdamage Xext Xfixes Xft Xi Xinerama
+WANTLIB += Xpm Xrandr Xrender atk-1.0 cairo dbus-1 fontconfig
+WANTLIB += freetype gdk-x11-2.0 gdk_pixbuf-2.0 gif gio-2.0 glib-2.0
+WANTLIB += gobject-2.0 gtk-x11-2.0 intl jpeg lcms2 pango-1.0 pangocairo-1.0
+WANTLIB += pangoft2-1.0 png rsvg-2 tiff xcb
 . elif ${FLAVOR} == "gtk3"
 CONFIGURE_ARGS+=   --with-x-toolkit=gtk3
 LIB_DEPENDS+=  x11/gtk+3
 WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
-WANTLIB += Xfixes Xft Xinerama Xpm Xrandr Xrender atk-1.0 cairo
-WANTLIB += cairo-gobject dbus-1 fontconfig freetype gconf-2 gdk-3
+WANTLIB += Xext Xfixes Xft Xinerama Xpm Xrandr Xrender atk-1.0
+WANTLIB += cairo cairo-gobject dbus-1 fontconfig freetype gdk-3
 WANTLIB += gdk_pixbuf-2.0 gif gio-2.0 glib-2.0 gobject-2.0 gtk-3
-WANTLIB += intl jpeg pango-1.0 pangocairo-1.0 png rsvg-2 tiff
-WANTLIB += xcb
+WANTLIB += intl jpeg lcms2 pango-1.0 pangocairo-1.0 png rsvg-2
+WANTLIB += tiff xcb
 . else
 ERRORS+=   "Fatal: 

Re: [pretest] emacs-26.0.90

[Jeremie Courreges-Anglas]

On Wed, Oct 25 2017, Jeremie Courreges-Anglas  wrote:
> On Sun, Oct 15 2017, Jeremie Courreges-Anglas  wrote:
>> Here's a diff for the first pretest in the emacs-26 release cycle.
>> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
>> is still used at build/unexec time).
>>
>> As usual build & test reports are welcome, especially if you use an arch
>> different from amd64 / arm, or if you use one of the graphical flavors.
>
> Thanks for the feedback so far, updated diff (sync with -current, use
> https instead of ftp).  I think we're going to stay with .xz tarballs
> for next releases, ~40M vs ~60M.
>
> The build is broken on armv7.  Tests on any arch you have at hand are
> still welcome.

The previous patch was messed up, sorry about that.


Index: Makefile
===
RCS file: /d/cvs/ports/editors/emacs/Makefile,v
retrieving revision 1.71
diff -u -p -r1.71 Makefile
--- Makefile19 Oct 2017 16:28:48 -  1.71
+++ Makefile26 Oct 2017 01:03:37 -
@@ -2,8 +2,7 @@
 
 COMMENT=   GNU editor: extensible, customizable, self-documenting
 
-VERSION=   25.3
-REVISION=  0
+VERSION=   26.0.90
 DISTNAME=  emacs-${VERSION}
 
 CATEGORIES=editors
@@ -17,7 +16,8 @@ PERMIT_PACKAGE_CDROM= Yes
 
 WANTLIB=   c m ncurses pthread gnutls xml2 z
 
-MASTER_SITES=  ${MASTER_SITE_GNU:=emacs/}
+MASTER_SITES=  https://alpha.gnu.org/gnu/emacs/pretest/
+EXTRACT_SUFX=  .tar.xz
 
 USE_GMAKE= Yes
 
@@ -47,46 +47,47 @@ LIB_DEPENDS=security/gnutls \
 .if ${FLAVOR} == "no_x11"
 CONFIGURE_ARGS+=   --without-x \
--without-dbus \
-   --without-gconf \
--without-gsettings \
-   --without-jpeg
+   --without-jpeg \
+   --without-lcms2
 .else
 LIB_DEPENDS+=  x11/dbus \
x11/gnome/librsvg \
-   devel/gconf2 \
graphics/jpeg \
graphics/png \
graphics/tiff \
graphics/giflib \
-   graphics/ImageMagick
+   graphics/ImageMagick \
+   graphics/lcms2
 RUN_DEPENDS+=  devel/desktop-file-utils \
devel/xdg-utils \
x11/gtk+3,-guic
 . if ${FLAVOR} == "athena"
 CONFIGURE_ARGS+=   --with-x-toolkit=athena
 LIB_DEPENDS+=  x11/Xaw3d
-WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 Xaw3d
-WANTLIB += Xext Xft Xinerama Xmu Xpm Xrandr Xrender Xt cairo dbus-1
-WANTLIB += fontconfig freetype gconf-2 gdk_pixbuf-2.0 gif gio-2.0
-WANTLIB += glib-2.0 gobject-2.0 intl jpeg png rsvg-2 tiff X11-xcb Xfixes xcb
+WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
+WANTLIB += Xaw3d Xext Xfixes Xft Xinerama Xmu Xpm Xrandr Xrender
+WANTLIB += Xt cairo dbus-1 fontconfig freetype gdk_pixbuf-2.0
+WANTLIB += gif gio-2.0 glib-2.0 gobject-2.0 intl jpeg lcms2 png
+WANTLIB += rsvg-2 tiff xcb
 . elif ${FLAVOR} == "gtk2"
 CONFIGURE_ARGS+=   --with-x-toolkit=gtk2
 LIB_DEPENDS+=  x11/gtk+2
-WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 Xcomposite
-WANTLIB += Xcursor Xdamage Xext Xfixes Xft Xi Xinerama Xpm Xrandr
-WANTLIB += Xrender atk-1.0 cairo dbus-1 fontconfig freetype gconf-2
-WANTLIB += gdk-x11-2.0 gdk_pixbuf-2.0 gif gio-2.0 glib-2.0 gobject-2.0
-WANTLIB += gtk-x11-2.0 intl jpeg pango-1.0 pangocairo-1.0 pangoft2-1.0
-WANTLIB += png rsvg-2 tiff X11-xcb xcb
+WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
+WANTLIB += Xcomposite Xcursor Xdamage Xext Xfixes Xft Xi Xinerama
+WANTLIB += Xpm Xrandr Xrender atk-1.0 cairo dbus-1 fontconfig
+WANTLIB += freetype gdk-x11-2.0 gdk_pixbuf-2.0 gif gio-2.0 glib-2.0
+WANTLIB += gobject-2.0 gtk-x11-2.0 intl jpeg lcms2 pango-1.0 pangocairo-1.0
+WANTLIB += pangoft2-1.0 png rsvg-2 tiff xcb
 . elif ${FLAVOR} == "gtk3"
 CONFIGURE_ARGS+=   --with-x-toolkit=gtk3
 LIB_DEPENDS+=  x11/gtk+3
-WANTLIB += ICE MagickCore-6.Q16 MagickWand-6.Q16 SM X11 X11-xcb
-WANTLIB += Xfixes Xft Xinerama Xpm Xrandr Xrender atk-1.0 cairo
+WANTLIB += Xext Xfixes Xft Xinerama Xpm Xrandr Xrender atk-1.0
+WANTLIB += cairo cairo-gobject dbus-1 fontconfig freetype gdk-3
 WANTLIB += cairo-gobject dbus-1 fontconfig freetype gconf-2 gdk-3
 WANTLIB += gdk_pixbuf-2.0 gif gio-2.0 glib-2.0 gobject-2.0 gtk-3
-WANTLIB += intl jpeg pango-1.0 pangocairo-1.0 png rsvg-2 tiff
-WANTLIB += xcb
+WANTLIB += intl jpeg lcms2 pango-1.0 pangocairo-1.0 png rsvg-2
+WANTLIB += tiff xcb
 . else
 ERRORS+=   "Fatal: Conflicting flavor: ${FLAVOR}"
 . endif
@@ -94,7 +95,6 @@ ERRORS+=  "Fatal: Conflicting flavor: ${
 
 post-install:
chmod -t 

Re: [pretest] emacs-26.0.90

[Jeremie Courreges-Anglas]

On Sun, Oct 15 2017, Jeremie Courreges-Anglas  wrote:
> Here's a diff for the first pretest in the emacs-26 release cycle.
> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
> is still used at build/unexec time).
>
> As usual build & test reports are welcome, especially if you use an arch
> different from amd64 / arm, or if you use one of the graphical flavors.

Thanks for the feedback so far, updated diff (sync with -current, use
https instead of ftp).  I think we're going to stay with .xz tarballs
for next releases, ~40M vs ~60M.

The build is broken on armv7.  Tests on any arch you have at hand are
still welcome.


Index: Makefile
===
RCS file: /d/cvs/ports/editors/emacs/Makefile,v
retrieving revision 1.71
diff -u -p -r1.71 Makefile
--- Makefile19 Oct 2017 16:28:48 -  1.71
+++ Makefile25 Oct 2017 16:56:58 -
@@ -2,8 +2,7 @@
 
 COMMENT=   GNU editor: extensible, customizable, self-documenting
 
-VERSION=   25.3
-REVISION=  0
+VERSION=   26.0.90
 DISTNAME=  emacs-${VERSION}
 
 CATEGORIES=editors
@@ -17,7 +16,8 @@ PERMIT_PACKAGE_CDROM= Yes
 
 WANTLIB=   c m ncurses pthread gnutls xml2 z
 
-MASTER_SITES=  ${MASTER_SITE_GNU:=emacs/}
+MASTER_SITES=  https://alpha.gnu.org/gnu/emacs/pretest/
+EXTRACT_SUFX=  .tar.xz
 
 USE_GMAKE= Yes
 
@@ -94,7 +94,6 @@ ERRORS+=  "Fatal: Conflicting flavor: ${
 
 post-install:
chmod -t ${PREFIX}/bin/emacs-${VERSION}
-   chmod -s ${PREFIX}/libexec/emacs/${VERSION}/${GCCARCH}/update-game-score
 .if ${FLAVOR:Mno_x11}
rm -Rf ${PREFIX}/share/applications/ ${PREFIX}/share/icons/
 .endif
Index: distinfo
===
RCS file: /d/cvs/ports/editors/emacs/distinfo,v
retrieving revision 1.10
diff -u -p -r1.10 distinfo
--- distinfo16 Sep 2017 16:42:27 -  1.10
+++ distinfo25 Oct 2017 02:41:27 -
@@ -1,2 +1,2 @@
-SHA256 (emacs-25.3.tar.gz) = 9yxqG0i2+6yiuZHu2AGWSiCKL4aGxwlAAT2ybNN5g8k=
-SIZE (emacs-25.3.tar.gz) = 62988771
+SHA256 (emacs-26.0.90.tar.xz) = 77JxJMuPPuupRy9PV3S12b9Ph/1NYCOq54Rp+1Znzyw=
+SIZE (emacs-26.0.90.tar.xz) = 43908428
Index: patches/patch-Makefile_in
===
RCS file: /d/cvs/ports/editors/emacs/patches/patch-Makefile_in,v
retrieving revision 1.3
diff -u -p -r1.3 patch-Makefile_in
--- patches/patch-Makefile_in   13 Oct 2016 15:39:32 -  1.3
+++ patches/patch-Makefile_in   25 Oct 2017 02:41:27 -
@@ -2,9 +2,10 @@ $OpenBSD: patch-Makefile_in,v 1.3 2016/1
 
 - don't compress man and info pages.
 
 Makefile.in.orig   Wed Jun 29 11:49:06 2016
-+++ Makefile.inTue Sep 27 15:22:19 2016
-@@ -664,7 +664,7 @@ install-info: info
+Index: Makefile.in
+--- Makefile.in.orig
 Makefile.in
+@@ -656,7 +656,7 @@ install-info: info
  for f in `ls $$elt $$elt-[1-9] $$elt-[1-9][0-9] 2>/dev/null`; do \
   (cd "$${thisdir}"; \
${INSTALL_DATA} ${srcdir}/info/$$f "$(DESTDIR)${infodir}/$$f"); 
\
@@ -13,7 +14,7 @@ $OpenBSD: patch-Makefile_in,v 1.3 2016/1
rm -f "$(DESTDIR)${infodir}/$$f.gz"; \
${GZIP_PROG} -9n "$(DESTDIR)${infodir}/$$f"; \
  done; \
-@@ -688,7 +688,7 @@ install-man:
+@@ -680,7 +680,7 @@ install-man:
  dest=`echo "$${page}" | sed -e 's/\.1$$//' -e '$(TRANSFORM)'`.1; \
  (cd "$${thisdir}"; \
   ${INSTALL_DATA} ${mansrcdir}/$${page} 
"$(DESTDIR)${man1dir}/$${dest}"); \
Index: patches/patch-configure
===
RCS file: /d/cvs/ports/editors/emacs/patches/patch-configure,v
retrieving revision 1.12
diff -u -p -r1.12 patch-configure
--- patches/patch-configure 1 May 2017 16:25:20 -   1.12
+++ patches/patch-configure 25 Oct 2017 02:41:27 -
@@ -1,7 +1,8 @@
 $OpenBSD: patch-configure,v 1.12 2017/05/01 16:25:20 jca Exp $
 configure.orig Thu Apr 20 17:50:43 2017
-+++ configure  Mon Apr 24 05:05:09 2017
-@@ -9530,7 +9530,14 @@ case "$opsys" in
+Index: configure
+--- configure.orig
 configure
+@@ -9219,7 +9219,14 @@ case "$opsys" in
openbsd)
 ## Han Boetes  says this is necessary,
 ## otherwise Emacs dumps core on elf systems.
Index: patches/patch-lisp_files_el
===
RCS file: /d/cvs/ports/editors/emacs/patches/patch-lisp_files_el,v
retrieving revision 1.2
diff -u -p -r1.2 patch-lisp_files_el
--- patches/patch-lisp_files_el 1 May 2017 16:25:20 -   1.2
+++ patches/patch-lisp_files_el 25 Oct 2017 02:41:27 -
@@ -2,9 +2,10 @@ $OpenBSD: patch-lisp_files_el,v 1.2 2017
 
 - open ports tree patches in diff-mode.
 
 lisp/files.el.orig Fri Feb  3 11:25:44 2017
-+++ lisp/files.el  

Re: [pretest] emacs-26.0.90

[Manuel Giraud]

Jeremie Courreges-Anglas  writes:

> Here's a diff for the first pretest in the emacs-26 release cycle.
> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
> is still used at build/unexec time).
>
> As usual build & test reports are welcome, especially if you use an arch
> different from amd64 / arm, or if you use one of the graphical
> flavors.

Using amd64/athena since two days now without issue.
-- 
Manuel Giraud

Re: [pretest] emacs-26.0.90

[Sunil Nimmagadda]

Jeremie Courreges-Anglas  writes:

> Here's a diff for the first pretest in the emacs-26 release cycle.
> Among other things, emacs now uses system malloc(3) at runtime (gmalloc
> is still used at build/unexec time).
>
> As usual build & test reports are welcome, especially if you use an arch
> different from amd64 / arm, or if you use one of the graphical flavors.

gtk3 flavor on amd64 builds/works fine for me, thanks.