[opensuse-factory] Emacs-27 startup borked
After the upgrade that comes with Emacs-27 things are not working too well. Emacs startup seems to be from source instead of byte-compiled file or a portable dump, suddenly uses the X11 backend instead of Xft or whatever Harfbuzz is supposed to use (with horrible font display that has no antialiasing, but still tries to scale the font) and Gnus doesn't even start and instead leaves this message: load-history-filename-element: Wrong type argument: stringp, (require . info) Similar messages are output for most of the other major modes I use, but I don't know where things have gone wrong. I found that I get things to work when starting Emacs as 'emacs -Q' (it then uses the Harfbuzz font backend as expected) and I can manually load my startup files later on and have it working (as far as I could quickly test it). I can use that as a temporary workaround for a few days probably. It does ostensibly not work when I start it with 'emacs -q -nsl --no-splash --name bork' (the latter is to avoid pulling in any X resources targeted at emacs). So it doesn't seem to be my own startup files, but rather sopmething pretty early in the startup sequence that -Q suppresses fortunately. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ Waldorf MIDI Implementation & additional documentation: http://Synth.Stromeko.net/Downloads.html#WaldorfDocs -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Achim Gratz writes:
I found that I get things to work when starting Emacs as 'emacs -Q' (it then uses the Harfbuzz font backend as expected) and I can manually load my startup files later on and have it working (as far as I could quickly test it). I can use that as a temporary workaround for a few days probably.
Gnus still doesn't workand fails with the same error message. So I need to use an older version of Emacs. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ SD adaptations for Waldorf Q V3.00R3 and Q+ V3.54R2: http://Synth.Stromeko.net/Downloads.html#WaldorfSDada -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
W dniu 19.08.2020 o 20:21, Achim Gratz pisze:
After the upgrade that comes with Emacs-27 things are not working too well. Emacs startup seems to be from source instead of byte-compiled file or a portable dump, suddenly uses the X11 backend instead of Xft or whatever Harfbuzz is supposed to use (with horrible font display that has no antialiasing, but still tries to scale the font)
Here's a short discussion: https://build.opensuse.org/package/show/editors/emacs The workaround that works for me is to use "emacs --no-x-resources"
On Wednesday, 19 August 2020 21:21:54 EEST Achim Gratz wrote:
suddenly uses the X11 backend instead of Xft or whatever Harfbuzz is supposed to use (with horrible font display that has no antialiasing, but still tries to scale the font)
This is a known issue, see https://bugzilla.suse.com/show_bug.cgi?id=1175372 for a fix.
Regards, Achim. -- Regards, Peter -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
auxsvr writes:
On Wednesday, 19 August 2020 21:21:54 EEST Achim Gratz wrote:
suddenly uses the X11 backend instead of Xft or whatever Harfbuzz is supposed to use (with horrible font display that has no antialiasing, but still tries to scale the font)
This is a known issue, see https://bugzilla.suse.com/show_bug.cgi?id=1175372 for a fix.
Well, I've had cleared the whole of xrdb to get rid of any of these things, but it seems that somewhere in the startup sequence the app defaults are pulled in anyway. OK, so I can work around this particular bug by either removing the starting emacs with --no-x-defaults. Done, but still no joy with Gnus… so I've compiled a fresh Emacs locally and installed it into /usr/local and that one works (after killing the Emacs app defaults) and does load it's pdumper file like it's supposed to instead of re-creating the whole thing (wrongly) from scratch at each start. This seems to be caused by starting Emacs as /usr/bin/emacs rather than emacs-gtk, so that particular fault is down to Emacs looking for the wrong pdump file. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ SD adaptation for Waldorf microQ V2.22R2: http://Synth.Stromeko.net/Downloads.html#WaldorfSDada -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2020/08/19 21:36:13 +0200, Achim Gratz wrote:
auxsvr writes:
On Wednesday, 19 August 2020 21:21:54 EEST Achim Gratz wrote:
suddenly uses the X11 backend instead of Xft or whatever Harfbuzz is supposed to use (with horrible font display that has no antialiasing, but still tries to scale the font)
This is a known issue, see https://bugzilla.suse.com/show_bug.cgi?id=1175372 for a fix.
Well, I've had cleared the whole of xrdb to get rid of any of these things, but it seems that somewhere in the startup sequence the app defaults are pulled in anyway. OK, so I can work around this particular bug by either removing the starting emacs with --no-x-defaults. Done, but still no joy with Gnus… so I've compiled a fresh Emacs locally and installed it into /usr/local and that one works (after killing the Emacs app defaults) and does load it's pdumper file like it's supposed to instead of re-creating the whole thing (wrongly) from scratch at each start. This seems to be caused by starting Emacs as /usr/bin/emacs rather than emacs-gtk, so that particular fault is down to Emacs looking for the wrong pdump file.
The /usr/bin/emacs script does start normally emacs-gtk and this binary does search for different emacs.pdmp versions werner/emacs> strace -f -o log emacs Warning: game dir '/var/games/emacs': Permission denied werner/emacs> grep pdmp log 25403 openat(AT_FDCWD, "/usr/bin/emacs-gtk.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25403 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25403 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs-gtk.pdmp", O_RDONLY|O_CLOEXEC) = 3 For the other toolkit this is: werner/emacs> strace -f -o log emacs Warning: game dir '/var/games/emacs': Permission denied werner/emacs> grep pdmp log 25450 openat(AT_FDCWD, "/usr/bin/emacs-x11.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25450 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25450 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs-x11.pdmp", O_RDONLY|O_CLOEXEC) = 3 For the three types of toolkits there are werner/emacs> rpm -qf /usr/lib/emacs/27.1/x86_64-suse-linux/emacs*.pdmp emacs-x11-27.1-0.x86_64 emacs-nox-27.1-0.x86_64 emacs-x11-27.1-0.x86_64 Btw: If you're running emacs as a server/daemon you should stop it before changing the toolkit or (re)installing a fresh version of emacs Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr
Dr. Werner Fink writes:
The /usr/bin/emacs script does start normally emacs-gtk and this binary does search for different emacs.pdmp versions
werner/emacs> strace -f -o log emacs Warning: game dir '/var/games/emacs': Permission denied werner/emacs> grep pdmp log 25403 openat(AT_FDCWD, "/usr/bin/emacs-gtk.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25403 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25403 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs-gtk.pdmp", O_RDONLY|O_CLOEXEC) = 3
Yes, but if you start it with a full path as /usr/bin/emacs it completely misses the boat: 3717 lstat("/usr", {st_mode=S_IFDIR|0755, st_size=124, ...}) = 0 3717 lstat("/usr/bin", {st_mode=S_IFDIR|0755, st_size=76850, ...}) = 0 3717 lstat("/usr/bin/emacs", {st_mode=S_IFREG|0755, st_size=4334, ...}) = 0 3717 openat(AT_FDCWD, "/usr/bin/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 3717 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 3717 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) And that's exactly what my shortcut did (because it would otherwise have picked up whatever test version I have currently installed as /usr/local/bin/emacs). I've changed the shortcut to directly run /usr/bin/emacs-gtk instead, but that should really not be necessary. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ Factory and User Sound Singles for Waldorf Q+, Q and microQ: http://Synth.Stromeko.net/Downloads.html#WaldorfSounds -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Achim Gratz wrote:
And that's exactly what my shortcut did (because it would otherwise have picked up whatever test version I have currently installed as /usr/local/bin/emacs). I've changed the shortcut to directly run /usr/bin/emacs-gtk instead, but that should really not be necessary.
Hmm, using emacs-gtk doesn't help here. I *do* have to specify -Q, else it wouldn't find any font I try to set via Options->Set Default Font... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Peter Suetterlin wrote:
Achim Gratz wrote:
And that's exactly what my shortcut did (because it would otherwise have picked up whatever test version I have currently installed as /usr/local/bin/emacs). I've changed the shortcut to directly run /usr/bin/emacs-gtk instead, but that should really not be necessary.
Hmm, using emacs-gtk doesn't help here. I *do* have to specify -Q, else it wouldn't find any font I try to set via Options->Set Default Font...
To follow-up myself, the font issue was a different thing, cured by a (temporary) Xresources entry (thanks Dan) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 2020/08/20 20:26:20 +0200, Achim Gratz wrote:
Dr. Werner Fink writes:
The /usr/bin/emacs script does start normally emacs-gtk and this binary does search for different emacs.pdmp versions
werner/emacs> strace -f -o log emacs Warning: game dir '/var/games/emacs': Permission denied werner/emacs> grep pdmp log 25403 openat(AT_FDCWD, "/usr/bin/emacs-gtk.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25403 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 25403 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs-gtk.pdmp", O_RDONLY|O_CLOEXEC) = 3
Yes, but if you start it with a full path as /usr/bin/emacs it completely misses the boat:
3717 lstat("/usr", {st_mode=S_IFDIR|0755, st_size=124, ...}) = 0 3717 lstat("/usr/bin", {st_mode=S_IFDIR|0755, st_size=76850, ...}) = 0 3717 lstat("/usr/bin/emacs", {st_mode=S_IFREG|0755, st_size=4334, ...}) = 0 3717 openat(AT_FDCWD, "/usr/bin/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 3717 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory) 3717 openat(AT_FDCWD, "/usr/lib/emacs/27.1/x86_64-suse-linux/emacs.pdmp", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
And that's exactly what my shortcut did (because it would otherwise have picked up whatever test version I have currently installed as /usr/local/bin/emacs). I've changed the shortcut to directly run /usr/bin/emacs-gtk instead, but that should really not be necessary.
Meanwhile fixed, thanks. Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr
participants (5)
-
Achim Gratz
-
Adam Mizerski
-
auxsvr
-
Dr. Werner Fink
-
Peter Suetterlin