April 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: crash in copy/paste 25 years, 7 months

Oliver Graf

Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: > I kinda assumed that DragMotionProto was some internal thing that > might be vaguely related to DnD. The DtDnD... stuff is probably only > the exported API. This might be true. I try to dig around in the libs... (oh, you did already) > > There is a misc-user in the most recent events, but from the description it > > seems as the paste came from another app. But this would not create a > > misc-user... > > Huh? where do you see the event type? The only info we have is the > backtrace. The 'recent keystrokes' stuff is not related to the crash! Oh, whoops, my fault. But on the other hand, it is a misc-user, that is created, cause your button-2 press will start a drag and the release will create a drop which again is this last misc-user. Did you do the experiment in a special buffer? I just tried the stunt in *scratch*. Perhaps some other binding is interfering with the experimental-dnd stuff? > evermann＠nats13:~$ nm --defined-only --print-file-name /usr/dt/lib/lib*.so /usr/openwin/lib/lib*.so | grep DragMotio > /usr/dt/lib/libXm.so:000b5c44 t DragMotion > /usr/dt/lib/libXm.so:000b593c t DragMotionProto > > isn't that motif? Yap. It is. And it must be a drog/drop. But why there is a motion? I really did not move, cause I use a trackball... perhaps you should try to twiggle it a bit. But not now. On Monday I can take a look again. > Always (well, two out of two). It's a bit tough for me to debug, the > only Sun with CDE I have access to, is back in Germany [1] and I don't > even have a debug build (or source) there anymore. Mhh, perhaps I should try it directly from the machine. Last time I started it using my linux box as display... but this is only a small chance... What XEmacs? Also Nikko? > There must be other people out there using Solaris with (or without) > CDE. Just click 3 times button1 and once button2, please. Yes, please try to reproduce everyone using solaris boxes. Regards, Oliver.

1 participants
0 comments

Re: crash in copy/paste 25 years, 7 months

Gunnar Evermann

Oliver Graf <ograf(a)rhein-zeitung.de> writes: > Oliver Graf <ograf(a)rhein-zeitung.de> writes: > > > > I get this on solaris2.6 in 20.4 when I click three times button1 to > > > select a line and then (without moving the mouse) click button2 to > > > yank the selection. I seem to remember having seen this ages ago, but > > > only on xemacsen compiled with support for CDE. > > > > > > Does the 'DragMotionProto' in the backtrace refer to DnD? Oliver, any > > > ideas? > > > > Mhhh, all CDE Dnd Functions carry names like DtDnd... there is none in the > > call stack. Long time since I last touched CDE. I kinda assumed that DragMotionProto was some internal thing that might be vaguely related to DnD. The DtDnD... stuff is probably only the exported API. > There is a misc-user in the most recent events, but from the description it > seems as the paste came from another app. But this would not create a > misc-user... Huh? where do you see the event type? The only info we have is the backtrace. The 'recent keystrokes' stuff is not related to the crash! Maybe somebody with the X sources lying around can check what's going on here. Alas: evermann＠nats13:~$ nm --defined-only --print-file-name /usr/dt/lib/lib*.so /usr/openwin/lib/lib*.so | grep DragMotio /usr/dt/lib/libXm.so:000b5c44 t DragMotion /usr/dt/lib/libXm.so:000b593c t DragMotionProto isn't that motif? > I can't reproduce it using your description with Nikko. It just starts a drag > for me, which is immediatly dropped and inserted. Does the bug appear always, > or only sometimes? Always (well, two out of two). It's a bit tough for me to debug, the only Sun with CDE I have access to, is back in Germany [1] and I don't even have a debug build (or source) there anymore. There must be other people out there using Solaris with (or without) CDE. Just click 3 times button1 and once button2, please. Gunnar Footnotes: [1] On the second try I just remotely f*cked up my local XServer -- I keep forgetting how dangerous breakpoints and crashes inside the Xlibs are. -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

1 participants
0 comments

Installation of 21.0 25 years, 7 months

Olaf Bachmann

Hey guys, I've been an enthusiastic and faithful xemacs user und apostel for many years now and always liked it very much. But now I have some complains w.r.t. version 21.0: I occoasionally check www.xemacs.org and today I found the announcment that v 21.0 is available. I downloaded it and compiled it -- fine. But then the trouble began. * I did not know of your new concept of packages -- so my "old" stuff in .emacs failed miserably. Very bad -- you should _SCREAM_ at us on the front page of www.xemacs.org and in all INSTALL/README files that we additionally need to install packages, before our "normal" xemacs stuff will work. * Ok -- so I figured that out. But then came the blow: In the info to packages, you make references to the node EFS, without distributing that node with the info files. You recommend that with EFS, package downloading and installation is easy -- but you do not put the EFS package in the core distribution. I tried all your methods for downloading packages and installing packages from within Xemacs, but always get the message: "Package-get entries not pgp signed: Coninue?" -- confirming, I get "Symbol's value as variable is void: allow-remote-paths" So, all that said, please either do not announce and make available v 21.0, or fix these simple problems. My users here at my site will lynch me, otherwise, if I install 21.0 for them on our machines. Thanks, Olaf

1 participants
0 comments

Re: crash in copy/paste 25 years, 7 months

Oliver Graf

Oliver Graf <ograf(a)rhein-zeitung.de> writes: > > I get this on solaris2.6 in 20.4 when I click three times button1 to > > select a line and then (without moving the mouse) click button2 to > > yank the selection. I seem to remember having seen this ages ago, but > > only on xemacsen compiled with support for CDE. > > > > Does the 'DragMotionProto' in the backtrace refer to DnD? Oliver, any > > ideas? > > Mhhh, all CDE Dnd Functions carry names like DtDnd... there is none in the > call stack. Long time since I last touched CDE. There is a misc-user in the most recent events, but from the description it seems as the paste came from another app. But this would not create a misc-user... I can't reproduce it using your description with Nikko. It just starts a drag for me, which is immediatly dropped and inserted. Does the bug appear always, or only sometimes? Regards, Oliver.

1 participants
0 comments

Re: symbol->obarray and Vpure_uninterned_symbol_table 25 years, 7 months

Hrvoje Niksic

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > Hrvoje Niksic <hniksic(a)srce.hr> writes: > > > Since we no longer have pure space, Vpure_uninterned_symbol_table is > > no longer needed. > > Sorry, but isn't this still needed? It isn't. Look at mark_object(). > We do not have pure space now but isn't that temporary? If the > portable dumper is ready don't we have read-only objects again and > thus the need for Vpure_uninterned_symbol_table or generalized > equivalent to store the symbols these read-only objects refer to? I believe read-only objects will use a bit different tactics for GC so the hack will not need to be reincarnated.

1 participants
0 comments

Re: new 21.2? 25 years, 7 months

Hrvoje Niksic

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > Even strings can point to other objects because they can have > properties. (Not sure whether there is actually code that does But of course. :-)

1 participants
0 comments

Re: new 21.2? 25 years, 7 months

Hrvoje Niksic

Colin Rafferty <craffert(a)ms.com> writes: > Has there been a 21.2 commit to CVS, but not announcement on > xemacs-beta? > > More specifically, does it remove purespace altogher? Cool hack. I > never was sure what we needed it for, since it is for sharing static > data between processes, and most OSs are smart about COW. COW is all nice and fine for one user, but I don't think it allows you to share the regions in question between users. Which is why pure space will return, but in a much healthier incarnation. Olivier, could you please outline what you intend to do here? I thought I understood it, but after a session with Kyle, I'm not so sure anymore. :-)

1 participants
0 comments

Re: any motif experts out there? 25 years, 7 months

Gunnar Evermann

Hironori Komaba <komaba(a)jpn.hp.com> writes: > When chars is set to "", chars should not be XtFree'ed. > I think this is a bug. You are right. I missed that completely. My excuse is that I looked at older versions all day and they didn't have this XtFree at all. :-) That also means that this can't be the cause of the crashes. Anyway, thanks for spotting this. I'll send a patch Gunnar -- Gunnar Evermann Speech, Vision & Robotics Group Engineering Department Cambridge University

1 participants
0 comments

Re: crash in copy/paste 25 years, 7 months

Oliver Graf

Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: > > #3 0xef6b63d4 in DragMotionProto () > > #4 0xef6b6984 in FinishAction () > > #5 0xef3a687c in HandleActions () > > #6 0xef3a6fb8 in HandleSimpleState () > > #7 0xef3a76b0 in _XtTranslateEvent () > > #8 0xef37cf60 in XtDispatchEventToWidget () > > #9 0xef37db9c in _XtDefaultDispatcher () > > #10 0xef37de44 in XtDispatchEvent () > > #11 0xef38bc54 in XtAppProcessEvent () > > #12 0x139a68 in emacs_Xt_event_pending_p (user_p=1) at event-Xt.c:2585 > > #13 0x9bd3c in detect_input_pending () at event-stream.c:945 > > I get this on solaris2.6 in 20.4 when I click three times button1 to > select a line and then (without moving the mouse) click button2 to > yank the selection. I seem to remember having seen this ages ago, but > only on xemacsen compiled with support for CDE. > > Does the 'DragMotionProto' in the backtrace refer to DnD? Oliver, any > ideas? Mhhh, all CDE Dnd Functions carry names like DtDnd... there is none in the call stack. Long time since I last touched CDE. But it seems possible that it is connected to Dnd, cause the button 2 should invoke a drag, if it is pressed within the selection. I will try to debug this later this day on my sun. > Speech, Vision & Robotics Group BTW: the Lego MindStorm is great... Regards, Oliver.

1 participants
0 comments

Re: The patch to end all redisplay problems. 25 years, 7 months

Jan Vroonhof

Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: > > * xemacs-20/src/window.c (struct window_config): Removed frame > > size members. > > (window_config_equal): No longer compare frame sizes. > > (Fset_window_configuration): Resize old top window to fit in the > > current frame directly, no longer use a fake frame resize. > > (Fcurrent_window_configuration): No longer save frame size > > > I have been running XEmacs with this patch for a couple of weeks and > just noticed, that resizing of minibuffers is broken (I think it > worked before). Somehow I missed that the old code restored also the size of the minibuffer. The patch needs to be enhanced with code that sets the minibuffer to the correct size before fitting in the rest. Jan

1 participants
0 comments

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

XEmacs-Beta April 1999