December 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: [PATCH] fix for Crash in mapcar1 25 years, 2 months

Hrvoje Niksic

Martin Buchholz <martin(a)xemacs.org> writes: > Jan> Why not do that? We can also do away with the caller calling > Jan> Flength when the results are not needed. Then we would only > Jan> loop over the list once and in addition the size of the list we > Jan> can loop over is not limited by the stack size limit. > > It's not obvious to me which implementation is faster. But yours introduces additional stack-allocation in the `mapc' case. > I actually have plans for optimizing map* in a bigger way. The > basic idea is to avoid repeating computations for each function > invocation. I think I had a patch that did that, but it was very non-general. In fact, it was one of my first patches to XEmacs C sources. A kind elderly lady persuaded me to give it up at the time. :-)

5 participants
6 comments

Re: Gutter redisplay too agressive. 25 years, 2 months

Jan Vroonhof

Andy Piper <andy(a)xemacs.org> writes: > >(save-current-window > > What's this --^^^^? I meant save-selected-window > Moreover I can't see any redisplay optimizations you refer to. Why not? > Sure redisplay doesn't kick in till events trigger, but this cannot > be true for widgets - you have to modify properties when you are > asked to and this will cause some flicker. Why can you do make changes to the internal structure and then have the visual representation sync up with them at with redisplay time? That is what happens with all the other stuff? > I wonder whether we could do some specific optimization > of this particular case (e.g. not call update-tab-in-gutter if we are in a > window-excursion)? I looked at that, but save-selected-window is bitch in that (just a macro). I still wonder why all that code uses select-window. Maybe there should be a set-window just like set-buffer vs switch-to-buffer? Jan

3 participants
5 comments

fill-region-as-paragraph / mule / pdump / please help 25 years, 2 months

Olivier Galibert

I'm stumped. fill-region-as-paragraph works perfectly (afaik) when not using pdump or not using mule. Only pdump+mule is broken. While I can mostly follow the lisp code when featurep 'mule returns nil, I'm overweirded when the kinsoku junk comes into play. Could someone more knowledgeable about lisp point me to the subr/variable/whatever that does not behave the way it should so that I can trace the problem at the C level? OG.

3 participants
3 comments

Re: check process state before accessing coding_stream (fix PR#1061) 25 years, 2 months

Gunnar Evermann

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > Gunnar Evermann <ge204(a)eng.cam.ac.uk> writes: > > > * process.c (Fprocess_input_coding_system): check whether process > > is still alive > > Approved. Recommened for 21.1. I think we should still apply the above patch to 21.1. > > process = get_process (process); > > + if (!PROCESS_LIVE_P (process)) > > + error ("Process %s is not active", > > + XSTRING_DATA (XPROCESS (process)->name)); > > + > > return decoding_stream_coding_system (XLSTREAM (XPROCESS (process)->coding_instream) ); > > You might want to turn this into a CHECK_PROCESS_LIVE(process) > macro[1], to be consistent with other such checks. yes, good point. I had actually looked for. I have now defined CHECK_LIVE_PROCESS and, to make all this consistent, a Lisp Function process-live-p (why wasn´t that defined before? Surely it is useful and nicer to use than process-status?). I also made the existing PROCESS_LIVE_P consistent with the other _LIVE_P macros and used it where possible (OK, so I got a bit carried away hacking process.c :-). There is one use of PROCESS_LIVE_P left (in process_send_signal) that could be replaced by CHECK_LIVE_PROCESS. I haven´t done that yet because the resulting error would be different (i.e. we wouldn´t print the process name anymore). any opinions on that? I´ll submit the crash test case for the test-suite separately. I would appreciate comments from people more familiar with process.c (Ben?). 1999-12-18 Gunnar Evermann <ge204(a)eng.cam.ac.uk> * process.h (PROCESS_LIVE_P): Modify to take a Lisp_Process instead of a Lisp_Object as argument to make it consistent with the other LIVE_P macros. (CHECK_LIVE_PROCESS): New macro. * process.c: Declare Qprocess_live_p. (Fprocess_live_p): New function. (create_process): Use PROCESS_LIVE_P. (read_process_output): Ditto. (set_process_filter): Ditto. (Fdelete_process): Ditto. (kill_buffer_processes): Ditto (Fprocess_input_coding_system): Check whether process is still alive (fix PR#1061). (Fprocess_output_coding_system): Ditto. (Fprocess_coding_system): Ditto. (Fset_process_input_coding_system): Ditto. (Fset_process_output_coding_system): Ditto. --- src/process.h.orig Sat Dec 18 16:39:02 1999 +++ src/process.h Sat Dec 18 16:39:11 1999 ＠＠ -45,7 +45,13 ＠＠ #define XSETPROCESS(x, p) XSETRECORD (x, p, process) #define PROCESSP(x) RECORDP (x, process) #define CHECK_PROCESS(x) CHECK_RECORD (x, process) -#define PROCESS_LIVE_P(x) (!NILP (XPROCESS(x)->pipe_instream)) +#define PROCESS_LIVE_P(x) (!NILP ((x)->pipe_instream)) + +#define CHECK_LIVE_PROCESS(x) do { \ + CHECK_PROCESS (x); \ + if (! PROCESS_LIVE_P (XPROCESS (x))) \ + dead_wrong_type_argument (Qprocess_live_p, (x)); \ +} while (0) #ifdef emacs --- src/process.c.orig Sat Dec 18 16:43:27 1999 +++ src/process.c Sat Dec 18 17:32:40 1999 ＠＠ -58,7 +58,7 ＠＠ #include "systty.h" #include "syswait.h" -Lisp_Object Qprocessp; +Lisp_Object Qprocessp, Qprocess_live_p; /* Process methods */ struct process_methods the_process_methods; ＠＠ -257,6 +257,14 ＠＠ return PROCESSP (obj) ? Qt : Qnil; } +DEFUN ("process-live-p", Fprocess_live_p, 1, 1, 0, /* +Return t if OBJECT is a process that is alive. +*/ + (obj)) +{ + return PROCESSP (obj) && PROCESS_LIVE_P (XPROCESS (obj)) ? Qt : Qnil; +} + DEFUN ("process-list", Fprocess_list, 0, 0, 0, /* Return a list of all processes. */ ＠＠ -510,7 +518,7 ＠＠ pid = PROCMETH (create_process, (p, argv, nargv, program, cur_dir)); p->pid = make_int (pid); - if (!NILP(p->pipe_instream)) + if (PROCESS_LIVE_P (p)) event_stream_select_process (p); } ＠＠ -823,7 +831,7 ＠＠ Really, the loop in execute_internal_event() should check itself for a process-filter change, like in status_notify(); but the struct Lisp_Process is not exported outside of this file. */ - if (NILP(p->pipe_instream)) + if (!PROCESS_LIVE_P (p)) return -1; /* already closed */ if (!NILP (p->filter) && (p->filter_does_read)) ＠＠ -1032,7 +1040,7 ＠＠ set_process_filter (Lisp_Object proc, Lisp_Object filter, int filter_does_read) { CHECK_PROCESS (proc); - if (PROCESS_LIVE_P (proc)) { + if (PROCESS_LIVE_P (XPROCESS (proc))) { if (EQ (filter, Qt)) event_stream_unselect_process (XPROCESS (proc)); else ＠＠ -1121,6 +1129,7 ＠＠ (process)) { process = get_process (process); + CHECK_LIVE_PROCESS (process); return decoding_stream_coding_system (XLSTREAM (XPROCESS (process)->coding_instream) ); } ＠＠ -1130,6 +1139,7 ＠＠ (process)) { process = get_process (process); + CHECK_LIVE_PROCESS (process); return encoding_stream_coding_system (XLSTREAM (XPROCESS (process)->coding_outstream)); } ＠＠ -1139,6 +1149,7 ＠＠ (process)) { process = get_process (process); + CHECK_LIVE_PROCESS (process); return Fcons (decoding_stream_coding_system (XLSTREAM (XPROCESS (process)->coding_instream)), encoding_stream_coding_system ＠＠ -1153,6 +1164,8 ＠＠ { codesys = Fget_coding_system (codesys); process = get_process (process); + CHECK_LIVE_PROCESS (process); + set_decoding_stream_coding_system (XLSTREAM (XPROCESS (process)->coding_instream), codesys); return Qnil; ＠＠ -1166,6 +1179,8 ＠＠ { codesys = Fget_coding_system (codesys); process = get_process (process); + CHECK_LIVE_PROCESS (process); + set_encoding_stream_coding_system (XLSTREAM (XPROCESS (process)->coding_outstream), codesys); return Qnil; ＠＠ -1527,7 +1542,7 ＠＠ if (network_connection_p (proc)) error ("Network connection %s is not a subprocess", XSTRING_DATA (XPROCESS(proc)->name)); - if (!PROCESS_LIVE_P (proc)) + if (!PROCESS_LIVE_P (XPROCESS (proc))) error ("Process %s is not active", XSTRING_DATA (XPROCESS(proc)->name)); ＠＠ -1883,7 +1898,7 ＠＠ p->tick++; process_tick++; } - else if (!NILP(p->pipe_instream)) + else if (PROCESS_LIVE_P (p)) { Fkill_process (proc, Qnil); /* Do this now, since remove_process will make sigchld_handler do nothing. */ ＠＠ -1915,7 +1930,7 ＠＠ { if (network_connection_p (proc)) Fdelete_process (proc); - else if (!NILP (XPROCESS (proc)->pipe_instream)) + else if (PROCESS_LIVE_P (XPROCESS (proc))) process_send_signal (proc, SIGHUP, 0, 1); } } ＠＠ -1978,6 +1993,7 ＠＠ syms_of_process (void) { defsymbol (&Qprocessp, "processp"); + defsymbol (&Qprocess_live_p, "process-live-p"); defsymbol (&Qrun, "run"); defsymbol (&Qstop, "stop"); defsymbol (&Qopen, "open"); ＠＠ -1991,6 +2007,7 ＠＠ #endif DEFSUBR (Fprocessp); + DEFSUBR (Fprocess_live_p); DEFSUBR (Fget_process); DEFSUBR (Fget_buffer_process); DEFSUBR (Fdelete_process);

5 participants
11 comments

zero size widget crashes continue 25 years, 2 months

Michael Harnois

This is not critical to me, since I kept a working copy of an older 21.2 beta to actually use ... but I haven't been able to use a current CVS beta for at least a month because, in gnus, they crash after no more than ten minutes with errors like these: xemacs: X Error of failed request: BadValue (integer parameter out of range for operation) Major opcode of failed request: 12 (X_ConfigureWindow) Value in failed request: 0x0 Serial number of failed request: 79626 Current serial number in output stream: 79659 Error: Widget *Article* has zero width and/or height -- Michael D. Harnois, Redeemer Lutheran Church, Washburn, IA mharnois(a)willinet.net aa0bt(a)aa0bt.ampr.org "Waiter, there's no fly in my soup!" - Kermit the Frog

4 participants
3 comments

Tabs, the gutter, `window-height', scrolling. (Was: Re: OK I give up) 25 years, 2 months

Karl M. Hegbloom

I imagine this must have been noticed by others, and is likely on someone's ToDo list anyway, but it can't hurt to point it out if I keep it brief... With the tabs displayed, when I push the page-down key, it doesn't put the line that was at the bottom on the top line like it used to. That line is off the top of the window now. I looked at `window-height', and see that it accounts for the modeline and possible scrollbar, but it does not (yet) account for the tab gutter, and needs to. I am not capable at this time of making that patch, or I would. I just plain don't understand things well enough yet to monkey with it.

2 participants
1 comments

Re: hyper-apropos-mouse-get-doc 25 years, 2 months

Adrian Aichner

>>>>> "Charles" == Charles G Waldman <cgw(a)fnal.gov> writes: Charles> Has anybody noticed that this function doesn't work in Charles> 21.1 (patch 8) "Bryce Canyon"? Charles> For instance if I type Charles> C-h a test RET Charles> the first entry in the list of "Functions and Macros" is Charles> add-latest-path Charles> If I click mouse-2 on this link, I get the right doc Charles> opened up in a bottom window. But within this doc Charles> window, the following of links does not seem to work. For me it works on both instances of `load-path', but not on `default-load-path'. I think this is because `default-load-path' does not exist as a variable. Adrian

3 participants
2 comments

Re: Bad addresses in wide reply. 25 years, 2 months

ShengHuo ZHU

>>>>> "Jérôme" == Jérôme Lacoste <lacostej(a)oslo.geco-prakla.slb.com> writes: Jérôme> Pterodactyl Gnus v0.96 Jérôme> XEmacs 21.1 (patch 3) "Acadia" [Lucid] (sparc-sun-solaris2.6) of Tue Jun 29 1999 on ardbeg Jérôme> 200 news.sinet.slb.com InterNetNews NNRP server INN 2.2.1 25-Aug-1999 ready (posting ok). Jérôme> I received a mail (cf RECEIVED MAIL section), with a header containing Jérôme> a lot of addresses. An error appeared when I tried to make a Wide Jérôme> reply (cf WIDE REPLY BUFFER section) : the mail server didn't accept Jérôme> the recipients (cf trace of SMTP session BUFFER). Jérôme> Gnus apparently processed badly the recipients, Jérôme> The error appears on the recipient Jérôme> "'Lacoste, Jerome (ORK)'" <lacostej(a)oslo.geco-prakla.slb.com>, Jérôme> which is shorten to "'Lacoste, Jérôme> Thus creating a parsing error (a '" is missing). [...] Jérôme> # start of WIDE REPLY BUFFER # Jérôme> ############################## Jérôme> To: "Eriksen, Kim Moore" <kim.moore.eriksen(a)folkehelsa.no> Jérôme> Cc: "'Ashton, Alan (ORK)'" <alan.ashton(a)unitor.com>, Jérôme> "'Bach-Gansmo, Tore (NRF)'" <tore.bach-gansmo(a)radiol.uu.se>, Jérôme> "'Bavridge, Nick'" <nickb(a)euroil.no>, Jérôme> "'Bjørkevoll, Johan (ORK)'" <johan.bjorkevoll(a)dnb.no>, Jérôme> "'Bressand, Jean Yves (ORK)'" <jyb(a)paellingsen.no>, Jérôme> "'Carriere, Stephane (ORK)'" <stephane.carriere(a)total.telemax.no>, Jérôme> "'Cooper, Colin (NRF)'" <ccolin(a)online.no>, Jérôme> "'Dalton, David'" <ddalton(a)se-as.no>, Jérôme> "'Davies, Jerry (NRF)'" <jdavies(a)eunet.no>, Jérôme> "'Deck, Bertrand (ORK)'" <bertrand.deck(a)hydro.com>, Jérôme> "'Demarest, Andre (ORK)'" <andre.v.demarest(a)no.pwcglobal.com>, Jérôme> "'Dinning, Jeff'" <jeff.dinning(a)get2net.dk>, Jérôme> "'Fogalele, Robin (ORK)'" <robin_fogalele(a)dmr.ca>, Jérôme> "'Gjertsen, Roar (NRF)'" <roar.gjertsen(a)iss-shipping.com>, Jérôme> "'Guvåg, Per Arne (ORK)'" <per.1395(a)online.no>, Jérôme> "'Hall, Keith (ORK)'" <keith.hall(a)winge.no>, Jérôme> "'Hodne, Per (NRF)'" <per.hodne(a)oslore.no>, Jérôme> "'Howells, J. Huw (ORK)'" <huw_jonathan_howells(a)dmr.com>, Jérôme> "'Johansson, Bradley (ORK)'" <bradley.johansson(a)avis.no>, Jérôme> "'Jones, Philip (ORK)'" <philip.jones1(a)chase.com>, Jérôme> "'Kittilsen, Jon (NRF)'" <jokittil(a)c2i.net>, "'Lacoste, "'Larsen, Jérôme> Øyvind (ORK)'" <oey-lar(a)online.no>, "'Laugerud, Jérôme> Eirik M. (ORK)'" <laueir(a)dph.no>, "'Lie, [...] It is a bug in `rmail-dont-reply-to'. The bug has been fixed in Gnu Emacs 20.4, while the bug still exists in XEmacs mail-lib (currently 1.24). (let ((rmail-dont-reply-to-names "test")) (require 'mail-utils) (rmail-dont-reply-to "a＠b, \"Last, First\" <test＠somewhere>")) => "a＠b, \"Last" It should be (as what does in Emacs 20.4) => "a＠b" -- Shenghuo

2 participants
1 comments

Re: Don't set coding-system in shell-mode 25 years, 2 months

Ben Wing

Yoshiki Hayashi wrote: > Ben Wing <ben(a)666.com> writes: > > > Martin Buchholz wrote: > > > > > I approve the patch, if this really works (it somewhat surprises me). > > > It did not work when I originally wrote this code. Does this also > > > work correctly for ttys? I mean, is the tty coding system properly > > > autodetected using determine_real_coding_system? > > > > > > Blue Sky: Can > > > > > > cat test.euc > > > cat test.jis > > > > > > possibly work `correctly'? > > Yes, because you can always distinguish that two > coding-system. When escape sequence `ESC [ $ B' is found, > it's most likely that coding-system is iso-2022-jp. When > two byte sequence has its MSB set, then it is probably > euc-jp. > > > > We know that the Latin equivalent > > > > > > cat test.iso8859-2 > > > cat test.iso8859-1 > > > > > > cannot. > > Because there's no particular characteristics in that > coding-system. It's just 8bit byte sequences. So it > depends on what language-environment you are in. If you > only use those two, it may be possible to disguinsh them > with some good heuristics. > > > Why not? Shell mode could reset things so as to force autodetection for the > > output of each command issued. Even your second example could work if those > > files had proper ISO2022 or X-Compound-Text or whatever escape sequences in > > them, or magic XEmacs coding-system: stuff, etc. to indicate the coding > > system. (Granted, not likely.) > > Here's a list of candidate to Mule-izing shell-mode. > 1. autodetect every time. > 2. autodetect only first time. > 3. set default coding-system and allow C-x RET c to alter > default value. > > Now I found doing this > cat test.ascii > cat test.sjis > will set process-output-coding-system to raw-text, I don't > think 2 is a good plan (my original one). > I think 3 is best as most user use only one coding-system. > I'll send patch to do this. Moved to xemacs-beta. I disagree. We always want to be automatic when possible -- i.e. option 1. We don't want to burden the user with extra work -- i.e. #3. > > > Besides that, it would be nice to have convenience way to > set coding-system of next command in shell-mode. Maybe > command to change process coding-system for permanently > would be nice, too. > > -- > Yoshiki Hayashi -- In order to save my hands, I am cutting back on my responses, especially to XEmacs-related mail. You _will_ get a response, but please be patient. If you need an immediate response and it is not apparent in your message, please say so. Thanks for your understanding.

3 participants
10 comments

nextaw lib & Xaw3d headers... broken again 25 years, 2 months

Giacomo Boffi

(insert (emacs-version)) XEmacs 21.2 (beta25) "Hephaestus" [Lucid] (i586-pc-linux) of Thu Dec 30 1999 on boffi95.stru.polimi.it i installed (xmkmf; make install) the latest nextaw libs AND headers the headers were put under /usr/X11R6/include/neXtaw, and all of libXaw and libXaw3d were left alone (no links) i then configured like this: % configure ... --with-athena=next --with-*=athena ... but it looks like configure decided to use the Xaw3d headers and the neXtaw libraries (it's recorded in config.status) of course, xemacs is broken in particular, it hangs nicely when i try to exit using the menubar having some modified buffer when i enter the dialog box a random letter is shown in the minibuffer, and while the dialog box responds to mouse events, xemacs doesn't, and when i finally kill the dialog box, that apparently does't communicate with xemacs, the whole app dies i include only the lisp backtrace, but if a stack dump is needed i will promptly do do you think i have to send my configure.log? next-command-event() byte-code("..." [popup-dialog-box dbox dispatch-event next-command-event] 2) # (catch result ...) # bind (dbox contents position) get-dialog-box-response(#<misc-user-event (call-interactively save-buffers-kill-emacs)> ("Save file /home/jim/a? " ("Yes" . act) ("No" . skip) ("Yes All" . automatic) ("No All" . exit) ("Cancel" . quit) ("Display The Current Buffer" . [#<compiled-function ... "...(11)" ... 2>]))) # (unwind-protect ...) # bind (next def elt char prompt map mouse-event user-keys actions no-cursor-in-echo-area action-alist help list actor prompter) map-y-or-n-p(#<compiled-function (buffer) "...(84)" [buffer-modified-p buffer buffer-base-buffer symbol-value-in-buffer save-buffers-skip buffer-file-name exiting buffer-offer-save buffer-size 0 arg t switch-buffer one-window-p delete-other-windows switched switch-to-buffer format "Save file %s? " "Save buffer %s? " buffer-name] 4> #<compiled-function (buffer) "...(8)" [buffer nil (save-buffer) (...)] 3> (#<buffer "a"> #<buffer "*scratch*"> #<buffer " *Minibuf-0*"> #<buffer " *Echo Area*"> #<buffer " *pixmap conversion*"> #<buffer " *Message-Log*"> #<buffer "*Group*">) ("buffer" "buffers" "save") ((?\^R #<compiled-function ... "...(11)" ... 2> "display the current buffer"))) # bind (switched switch-buffer exiting arg) save-some-buffers-1(nil t t) # (unwind-protect ...) # (unwind-protect ...) # bind (exiting arg) save-some-buffers(nil t) # bind (arg) save-buffers-kill-emacs(nil) # bind (command-debug-status) call-interactively(save-buffers-kill-emacs) # (condition-case ... . error) # (catch top-level ...)

1 participants
1 comments

2025

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 December 1999