November 1999 - XEmacs-Beta - XEmacs Mailing List Archives

Re: [21.2] LDAP add and delete functions + interrupt fixes 25 years

William M. Perry

Ben Wing <ben(a)666.com> writes: > Why don't you use a more standard API? LDAP is basically a database, so > why don't you model the API after the other database interfaces in > XEmacs? The other databases have no heirarchy to them, and they are all file based. You could pervert the 'file' argument to open-database to be a database or hostname. The other main problem is that get-database currently expects there to be only one value per 'key'. To merge the LDAP code, you would have to do a major overhaul of the existing database code, and I'm not sure you would be able to get the full power of LDAP without making the 'normal' databases pretty difficult to use. -Bill P.

2 participants
1 comments

Re: Swallowing windows in X... 25 years

Jan Vroonhof

wmperry(a)aventail.com (William M. Perry) writes: > Anybody with more X experience want to comment? Basically, I changed > x_subwindow_instantiate() to do: > > [ugly "first try interface" deleted to protect minors] > > XReparentWindow (dpy, win, IMAGE_INSTANCE_X_CLIPWINDOW (ii), 0, 0); > XChangeWindowAttributes (dpy, win, valueMask, &xswa); > XSetWindowColormap (dpy, win, DefaultColormapOfScreen (xs)); I am not an X expert, but I can stand on the shoulders of Rob Nation (from the FVWM 2.2 sources:) XReparentWindow(dpy,Buttons[button].IconWin, main_win, j*ButtonWidth+2, i*ButtonHeight+2); XMapWindow(dpy,Buttons[button].IconWin); XSelectInput(dpy,(Window)body[0], PropertyChangeMask|StructureNotifyMask); Note that the FVWM swallow routine also looks up the WM hints to determine a natural size. I can mail you the complete routine if you want. Note that the gnome sources do something very weird when deleting windows. They reparent back to root and then delete .... Jan P.S. With all this X hackery going on it is becoming reasonable to define set-x-window-property, set-x-window-property-change-callback etc for elisp. P.P.S. Please submit a comprehensive GPM patch to xemacs-patches. I have lost track. I will apply it to 21.2 then.

2 participants
1 comments

Re: Ethiopic in 21.1.7++ 25 years

Vin Shelton

I did not have mule-base installed when I ran my very simple test of attempting to set each language environment one at a time. None of the other language environments gave me a problem, BTW. - vin >>>>> On 09 Nov 1999, Yoshiki Hayashi <t90553(a)m.ecc.u-tokyo.ac.jp> said: Yoshiki> "Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> writes: >> >>>>> "Jan" == Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: >> Jan> Vin Shelton <acs(a)xemacs.org> writes: >> >> Symbol's function definition is void: >> >> setup-ethiopic-environment-internal >> >> >> >> I'd like to get this working before I release 21.1.8 (tonight, >> >> I hope). Can anyone offer a patch? >> >> This looks like the same thing that was happening with Japanese and >> some other languages. The FSF changed the language environemnt API, >> and that got ported to a couple of language environments but not >> the rest. :-( Yoshiki> Like Jan said, it must not be Symbol's function definition Yoshiki> is void error since it is defined at ethio-util.el Maybe Vin Yoshiki> hasn't installed mule-base? However, it requires Yoshiki> exit-function handling to be implemented in Yoshiki> setup-language-environment so it is problematic to use code Yoshiki> in ethio-util.el (it defines f2, f3 and other keys so after Yoshiki> setting up Ethiopic language environment in XEmacs 21.1, the Yoshiki> original definition of those keys won't be recovered). Jan> I tried making the obvious path (namely copying Jan> 'setup-ethiopic-environment-internal from an old mule base Jan> into the core), however then I stopped. auto-autoloads.el Jan> from mule-base contains >> >> Ethiopic is one of the things that got ported from FSF and ignored, I >> think disabling it is the best route until somebody is willing to take >> a good look at it. Yoshiki> I think so, too. Until someone who knows Ethiopic is going Yoshiki> to use and test it, removing Ethiopic support would not Yoshiki> hurt. >> That is my suggestion for any non-ISO-8859, non-Asian-ideographic >> language that breaks, unfortunately. I think a real solution to this, >> one that is robust to ++patchlevel, even, will have to wait until we >> define real APIs for Mule and convince porters from FSF to stick to >> them. Yoshiki> Yes, we need APIs.

1 participants
0 comments

Re: [21.2] LDAP add and delete functions + interrupt fixes 25 years

Ben Wing

Why don't you use a more standard API? LDAP is basically a database, so why don't you model the API after the other database interfaces in XEmacs? ben Oscar Figueiredo wrote: > LDAP calls should now be more robust with regard to interruptions. New > functions for adding and deleting records from an LDAP server. Record > modification is on its way. > > Oscar > > 1999-11-14 Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch> > > * eldap.h(Fldap_add_internal, Fldap_delete_internal): New > functions > * eldap.c (Fldap_add_internal, Fldap_delete_internal): New > functions > (Fldap_open): Set the `LDAP_OPT_RESTART' option to avoid LDAP > calls being interrupted by XEmacs signals > (Fldap_search_internal): Remove useless calls to > `slow_down_interrupts' and `speed_up_interrupts' > > 1999-11-14 Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch> > > * ldap.el (toplevel): Signal error when XEmacs has been compiled > without LDAP > (ldap-add, ldap-delete): New defuns

1 participants
0 comments

Re: latest GPM patch... 25 years

William M. Perry

Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: > wmperry(a)aventail.com (William M. Perry) writes: > > > Index: lisp/gpm.el > > =================================================================== > > --- lisp/gpm.el Tue May 5 15:32:27 1998 > > +++ lisp/gpm.el Mon Nov 8 11:22:39 1999 > > Thanks, > > I'll have one more look at them and then apply them to 21.2. I assume > the ChangeLogs you posted before are still current. Yes. -bp

1 participants
0 comments

Re: Cosmetic fixes 25 years

Gunnar Evermann

Hrvoje "Martin" Niksic <hniksic(a)iskon.hr> writes: > Martin alert. Except that I'm not Martin. > [snip patches] Hey, where are the ChangeLogs? Sorry, couldn't resist -- I'm sure that's the response Martin would have expected :-) Gunnar

1 participants
0 comments

Cosmetic fixes 25 years

Hrvoje Niksic

Martin alert. Except that I'm not Martin. Index: src/lread.c =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/src/lread.c,v retrieving revision 1.36.2.13 diff -u -r1.36.2.13 lread.c --- lread.c 1999/10/10 12:40:11 1.36.2.13 +++ lread.c 1999/11/14 18:04:39 ＠＠ -903,17 +903,18 ＠＠ nil, a list, or a string (for backward compatibility), with the following semantics: - a) nil - no suffix, just search for file name intact (semantically - different from "empty suffix list") + a) nil - no suffix, just search for file name intact + (semantically different from "empty suffix list", which + would be meaningless.) b) list - list of suffixes to append to file name. Each of these must be a string. c) string - colon-separated suffixes to append to file name (backward compatibility). - All of this got hairy, so I decided to use write a mapper. Calling - a function for each suffix shouldn't slow things down, since - locate_file is rarely call with enough suffixes for it to make a - difference. */ + All of this got hairy, so I decided to use a mapper. Calling a + function for each suffix shouldn't slow things down, since + locate_file is rarely called with enough suffixes for funcalls to + make any difference. */ /* Map FUN over SUFFIXES, as described above. FUN will be called with a char * containing the current file name, and ARG. Mapping stops when Index: etc/NEWS =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/etc/NEWS,v retrieving revision 1.58.2.23 diff -u -r1.58.2.23 NEWS --- NEWS 1999/09/26 19:46:46 1.58.2.23 +++ NEWS 1999/11/14 18:04:40 ＠＠ -236,15 +236,21 ＠＠ colon-separated string. MODE is now a symbol or a list of symbols (symbols `exists', `executable', `writable', and `readable' are supported) instead of an integer code. See the documentation for -details. +details. Of course, the old form is still accepted for backward +compatibility. -Of course, the old form is still accepted for backward compatibility. +Several bugs in locate-file have been fixed, most notably its failure +to call expand-file-name on elements of PATH-LIST. Because of that +elements of load-path of the form "~/..." used to not work. +locate-file is now guaranteed to expand files during its course of +operation. ** `translate-region' has been improved in several ways. Its TABLE argument used to be a 256-character string. In addition to this, it -can now also be a vector or a char-table (which is useful for Mule.) -If TABLE a vector or a generic char-table, you can map characters to -strings instead of to other characters. For instance: +can now also be a vector or a char-table, which makes the function +useful for Mule, which it wasn't. If TABLE a vector or a generic +char-table, you can map characters to strings instead of to other +characters. For instance: (let ((table (make-char-table 'generic))) (put-char-table ?a "the letter a" table)

1 participants
0 comments

Re: stop_interrupts not working (was: Re: I need help from a signal expert !) 25 years

Jan Vroonhof

Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch> writes: > I made the same test with xemacs -nw and the problem does not show up in that > case. The guilty device(s) seem therefore to be X-related. I don't think you can include that so easily. terminal devices produce much less I/O. > I have no idea how to trace the SIGIOs back to the device(s) that generated > them. Call select? Jan

1 participants
0 comments

Re: latest GPM patch... 25 years

Jan Vroonhof

wmperry(a)aventail.com (William M. Perry) writes: > Index: lisp/gpm.el > =================================================================== > --- lisp/gpm.el Tue May 5 15:32:27 1998 > +++ lisp/gpm.el Mon Nov 8 11:22:39 1999 Thanks, I'll have one more look at them and then apply them to 21.2. I assume the ChangeLogs you posted before are still current. Jan

1 participants
0 comments

Re: I need help from a signal expert ! 25 years

Oscar Figueiredo

>>>>> "Jan" == Jan Vroonhof <vroonhof(a)math.ethz.ch> writes: Jan> Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch> writes: >> surrounded by {slow_down,speed_up}_interrupts. I even tried >> `stop_interrupts> and `start_interrupts' with no success. I can't >> understand why it works for searches (which also call `ldap_result') and >> not for addition and deletion. Jan> That could just be some timing issue. It could also be that there is a Jan> file descriptor involved that unrequest_sigio doesn't know about. >> I'd appreciate very much some help from someone knowledgeable in >> this area. Jan> Could you check with strace or truss which signal is interrupting the Jan> syscal? >From what I saw in DDD, it's a SIGIO. I have been investigating the problem a little further since my last post and it appears that `slow_down_interrupts' does not disable SIGIO correctly at least on RedHat 6.0 Linux. Adding a pair of EMACS_{UN}BLOCK_SIGNAL (SIGIO); around calls to `ldap_result' solves the problem. Therefore I believe `unrequest_sigio' (called from `slow_down_interrupts') is not doing its job correctly. Oscar

2 participants
1 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