November 1999 - XEmacs-Beta - XEmacs Mailing List Archives

William M. Perry

Well, I have a halfway working version of the GPM code using connect_to_file_descriptor() - it gets and reads events. And you can turn gpm-mode on and off and have it do the right thing. The only problem is that I am using Fdispatch_event() after creating an event and filling in the structures. Like this: ----- fake_event = Fmake_event (Qnil, Qnil); event = XEVENT(fake_event); event->timestamp = 0; event->channel = Fselected_frame (Qnil); /* CONSOLE_SELECTED_FRAME (con); */ if (ev.modifiers & 1) modifiers |= MOD_SHIFT; if (ev.modifiers & 2) modifiers |= MOD_META; if (ev.modifiers & 4) modifiers |= MOD_CONTROL; if (ev.modifiers & 8) modifiers |= MOD_META; [...] /* Handle the event */ Fdispatch_event (fake_event); Fdeallocate_event (fake_event); GPM_DRAWPOINTER(&ev); return (Qzero); ----- For one thing, you never really get to see the GPM pointer except for a lot of flickering, I think because redisplay happens immediately after the 'process output' is handled. Also, I get all sorts of weird errors like: (3) (error/warning) Error in process filter: (error Selection aborted) (4) (error/warning) Error in process filter: (wrong-type-argument integerp #<INTERNAL OBJECT (XEmacs bug?) (symbol-value-forward type 13) 0x822037c>) I think the selection aborted is because of the process events getting there when it thinks a mouse button is depressed, but I don't know for sure. Any ideas? -Bill P.

1 participants
0 comments

Re: compiler warnings mess 25 years

Jan Vroonhof

Stef Epardaud <stef(a)lunatech.com> writes: > labyrinth stephane 18:31 ...src/xemacs > gcc --version > egcs-2.91.66 > > that's not 1.1, right? I don't know. What I do know is that gcc 2.95 compiles 21.1.7 with almost no warnings at all and so does the Solaris compiler. As to which things to change and what not. Use your own judgement. If you get the feeling that you are only trying to work around the compiler being stupid then don't. If the code wasn't clear and the change has zero or almost no cost then do. However, Martin compiles XEmacs with a C++ compiler from time to time and then fixes all the warnings for that (a C++ compiler is allowed to me much more picky). Jan

2 participants
1 comments

Re: compiler warnings mess 25 years

Jan Vroonhof

Stef Epardaud <stef(a)lunatech.com> writes: > and in the end i even wondered if my 10-20 warnings fix were valid > because the ones remaining seemd unjustified. > what should i do, or is it normal to get dozens of : > window.c: In function `window_loop': > window.c:2409: warning: `w' might be used uninitialized in this function > window.c:2411: warning: `next_window' might be used uninitialized in > this function Are you using egcs 1.1 ? It has a known bug that causes it to warn far too much. Jan

2 participants
1 comments

Re: ViaVoice support (was Re: Correct format for patch attachments) 25 years

Jan Vroonhof

wmperry(a)aventail.com (William M. Perry) writes: > Hmmm... interesting. So when you connect_to_file_descriptor(), the data > doesn't actually get _read_ if you have a process filter set up? Would it > make sense to make the GPM code do this as well then? Maybe. The only thing is that then you don't get the console association. That is probably wrong anyway because the GPM library uses only one file descriptor to catch events for all Linux VT's and these are different consoles (they really shouldn't be but there is no way we can figure that out). Jan

1 participants
0 comments

Re: compiler warnings mess 25 years

Andreas Jaeger

>>>>> Stef Epardaud writes: > hi, > i decided to rebuild on cvs today, and although it did build, the > compilation gave me loads of warnings, and since i was really bored, i > took a look at them. > while i fixed (what i call 'fix' ... ;) some of them and produce some > patches, the dozens others remain and i cannot fix them because i don;'t > understand them, they seem plain wrong. > those i could fix were some with unused variables (forgotten somehow) gcc isn't very good in diagnosing if a variable is really used unitialized or not. Too often the warning is wrong. That's a known 'feature' of gcc:-(. Andreas -- Andreas Jaeger SuSE Labs aj(a)suse.de private aj(a)arthur.rhein-neckar.de

1 participants
0 comments

Re: ViaVoice support (was Re: Correct format for patch attachments) 25 years

William M. Perry

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > wmperry(a)aventail.com (William M. Perry) writes: > > > SmOpen(). I was planning on just throwing this into the event loop. I'm > > not sure if we can currently do this from the module API or not. > > I cannot think why not. Note that the tooltalk stuff which really > should be a module already does things on the lines of what you want. Hmmm... interesting. So when you connect_to_file_descriptor(), the data doesn't actually get _read_ if you have a process filter set up? Would it make sense to make the GPM code do this as well then? Would definitely make it cleaner, and the GPM code could become a module, as could the speech stuff. -Bill P.

1 participants
0 comments

Re: ViaVoice support (was Re: Correct format for patch attachments) 25 years

Jan Vroonhof

wmperry(a)aventail.com (William M. Perry) writes: > SmOpen(). I was planning on just throwing this into the event loop. I'm > not sure if we can currently do this from the module API or not. I cannot think why not. Note that the tooltalk stuff which really should be a module already does things on the lines of what you want. Jan

1 participants
0 comments

Stupid question: How to get Backspace working in TTY mode? 25 years

Michael Sperber

I really felt stupid when someone asked me how to make Backspace delete backwards in TTY mode, and I didn't know. This seems to be FAQ material, but isn't in the FAQ. -- Cheers =8-} Mike Friede, Völkerverständigung und überhaupt blabla

3 participants
2 comments

patch for portable dumper w/o debug 25 years

William M. Perry

Since debug.o is not linked unless you compile with debugging info, reinit_vars_of_debug() is undefined. This fixes it for me. I know I know, debug symbols good for portable dumper, but I configured that way by mistake and turned this up. :) -Bill P. Index: emacs.c =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/src/emacs.c,v retrieving revision 1.82.2.31 diff -c -w -r1.82.2.31 emacs.c *** emacs.c 1999/10/30 19:57:32 1.82.2.31 --- emacs.c 1999/11/03 11:56:21 *************** *** 1662,1668 **** --- 1662,1670 ---- reinit_vars_of_buffer (); reinit_vars_of_console (); + #ifdef DEBUG_XEMACS reinit_vars_of_debug (); + #endif reinit_vars_of_device (); reinit_vars_of_eval (); #ifdef HAVE_X_WINDOWS

1 participants
0 comments

ViaVoice support (was Re: Correct format for patch attachments) 25 years

William M. Perry

Ben Wing <ben(a)666.com> writes: > A subprocess would be preferred. (Until someone writes a foreign-function > interface so you can call arbitrary C functions from Lisp ...) Well, the reason I didn't start writing it as a module is that it is going to hook into the Xt or unixoid event loops, since you can get a file descriptor that gets tickled whenever new data is available. This is done with the 'SmNexternalNotifier' attribute in the SmArg list sent to SmOpen(). I was planning on just throwing this into the event loop. I'm not sure if we can currently do this from the module API or not. The route I would take with a subprocess is to use (make-event ...) with data from the subprocess. -Bill P.

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 November 1999