June 1998 - XEmacs-Beta - XEmacs Mailing List Archives

gnuclient -batch only prints first line of output 26 years, 5 months

Samuel Mikes

--Ruby_Ridge-+qJ6SaWl2p Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I was playing with gnuclient and discovered that gnuclient only writes out the first line of output: $ gnuclient -batch -eval '(print "foobar\nbletch\n")' ==> foobar I tracked it down to the use of read_line() in the output-producing cases of gnuclient; read_line only reads up to the first "\n" and stops, even if there's more output. The attached patch provides a fix which incidentally allows an aribtrary amount of data (not just GSERV_BUFSZ) to be received by gnuclient. --Ruby_Ridge-+qJ6SaWl2p Content-Type: text/plain Content-Description: Patch to gnuclient.c Content-Disposition: attachment; filename="gnuclient.patch" Content-Transfer-Encoding: 7bit --- gnuclient.c.orig Sun May 24 07:40:19 1998 +++ gnuclient.c Fri Jun 5 20:29:54 1998 ＠＠ -268,6 +268,42 ＠＠ return new; } +/* read possibly multiline result from -batch */ +char* +read_result (int s) +{ + char *temp=NULL; + char *result=NULL; + int length; + long offset=0; + long result_length=GSERV_BUFSZ; + + result = malloc (result_length+1); + + while ((length=read(s,result+offset,1)>0) && + result[offset] != EOT_CHR) + { + offset += length; + if (offset >= result_length) + /* copy and reallocate */ + { + result[offset] = '\0'; + temp = strdup (result); + + result_length += GSERV_BUFSZ; + result = realloc (result, result_length+1); + + strcpy (result, temp); + + free (temp); + } /* if */ + } /* while */ + + result[offset] = '\0'; + + return result; +} + int main (int argc, char *argv[]) { ＠＠ -303,7 +339,7 ＠＠ #endif /* SYSV_IPC */ char *tty = NULL; char buffer[GSERV_BUFSZ + 1]; /* buffer to read pid */ - char result[GSERV_BUFSZ + 1]; + char *result = NULL; int i; #ifdef INTERNET_DOMAIN_SOCKETS ＠＠ -441,7 +477,6 ＠＠ exit (1); } - *result = '\0'; if (eval_function || eval_form || load_library) { #if defined(INTERNET_DOMAIN_SOCKETS) ＠＠ -469,11 +504,7 ＠＠ } send_string (s, "))"); send_string (s, EOT_STR); - if (read_line (s, result) == 0) - { - fprintf (stderr, "%s: Could not read\n", progname); - exit (1); - } + result = read_result(s); } /* eval_function || eval_form || load_library */ else if (batch) { ＠＠ -495,11 +526,7 ＠＠ } send_string(s,"))"); send_string (s, EOT_STR); - if (read_line (s, result) == 0) - { - fprintf (stderr, "%s: Could not read\n", progname); - exit (1); - } + result = read_result (s); } if (!batch) ＠＠ -520,12 +547,14 ＠＠ send_string (s, "(gnuserv-eval '(emacs-pid))"); send_string (s, EOT_STR); + /* SOM: throwing away output here */ if (read_line (s, buffer) == 0) { fprintf (stderr, "%s: Could not establish Emacs procces id\n", progname); exit (1); } + /* Don't do disconnect_from_server becasue we have already read data, and disconnect doesn't do anything else. */ #ifndef INTERNET_DOMAIN_SOCKETS ＠＠ -648,8 +677,13 ＠＠ #endif /* !SYSV_IPC */ } /* not batch */ - if (batch && !quick) - printf ("%s\n", result); + if (batch && !quick && result) + printf ("%s", result); + + if (result != NULL) + { + free (result); + } return 0; --Ruby_Ridge-+qJ6SaWl2p Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cheers, -- Sam Mikes smikes(a)alumni.hmc.edu --Ruby_Ridge-+qJ6SaWl2p--

1 participants
0 comments

Re: 21.0-b34 ldap crash on Solaris 2.5.1 26 years, 5 months

Raymond Toy

Hrvoje Niksic <hniksic(a)srce.hr> writes: > Raymond, I assume you are building with Mule? If so, then please see > if this (untested) patch fixes the problem. > Applied the patch and rebuilt. Works just fine; no crashes on my example. Thanks! Ray

1 participants
0 comments

ediff-use-faces,highlight-all-diffs default nil? 26 years, 5 months

Greg Klanderman

Was it intentional that with the ediff 1.08 XEmacs package that highlighting is now off by default? (I do get some fine highlighting nevertheless but the text markers are UGLY). The relavant diffs seem to be (ediff-init.el): -(ediff-defvar-local ediff-use-faces t +(ediff-defvar-local ediff-use-faces nil "") +(defcustom ediff-use-faces t "If t, differences are highlighted using faces, if device supports faces. -(ediff-defvar-local ediff-highlight-all-diffs t +(ediff-defvar-local ediff-highlight-all-diffs nil "") +(defcustom ediff-highlight-all-diffs t "If nil, only the selected differences are highlighted. greg

1 participants
0 comments

Re: 21.0-b34 ldap crash on Solaris 2.5.1 26 years, 5 months

Hrvoje Niksic

Raymond, I assume you are building with Mule? If so, then please see if this (untested) patch fixes the problem. 1998-06-05 Hrvoje Niksic <hniksic(a)srce.hr> * eldap.c (Fldap_search_internal): Use build_ext_string instead of build_string to avoid crashes under Mule. --- eldap.c.orig Fri Jun 5 18:05:24 1998 +++ eldap.c Fri Jun 5 18:06:47 1998 ＠＠ -505,13 +505,13 ＠＠ a != NULL; a= ldap_next_attribute (ld, e, ptr) ) { - list = Fcons (build_string (a), Qnil); + list = Fcons (build_ext_string (a, FORMAT_OS), Qnil); unwind.vals = ldap_get_values (ld, e, a); if (unwind.vals != NULL) { for (i = 0; unwind.vals[i] != NULL; i++) { - list = Fcons (build_string (unwind.vals[i]), + list = Fcons (build_ext_string (unwind.vals[i], FORMAT_OS), list); } } -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- Which is worse: ignorance or apathy? Who knows? Who cares?

1 participants
0 comments

Re: infinite loading loop in eudc 26 years, 5 months

Oscar Figueiredo

>>>>> "Raymond" == Raymond Toy <toy(a)rtp.ericsson.se> writes: Raymond> Completely reproducible for me. Raymond> 1. Remove your ~/.eudc-options file. Raymond> 2. Use customize to set some eudc option and save it. My options Raymond> file looks like Raymond> ;; This file was automatically generated by eudc.el Raymond> (eudc-set-server "mbb1.ericsson.se" 'ldap t) Raymond> (setq eudc-server-hotlist 'nil) Raymond> (provide 'eudc-options-file) Raymond> 3. Quit and start up another xemacs -vanilla Raymond> 4. Click on Tools->Directory->Load Hotlist of Servers Raymond> I get an infinite loop continually loading ~/.eudc-options. Don't Raymond> know the cause but should be easy to find. Perhaps the features are Raymond> set in time to prevent repeated loading. The provide statement should be first in the file. I will fix it for the next release of eudc. Oscar

1 participants
0 comments

Re: [patch] correct fix (Was: Fix for Re: Bug (I think) limit on popup menu strings) 26 years, 5 months

Hrvoje Niksic

Colin Rafferty <craffert(a)ms.com> writes: > Here is a patch that also works, but sets no limit on the length of > the string. Use alloca instead of malloc, and that patch can go in. Let's fight arbitrart limits! -- Hrvoje Niksic <hniksic(a)srce.hr> | Student at FER Zagreb, Croatia --------------------------------+-------------------------------- "Beware of bugs in the above code; I have only proved it correct, not tried it." -- Donald Knuth

1 participants
0 comments

[patch] correct fix (Was: Fix for Re: Bug (I think) limit on popup menu strings) 26 years, 5 months

Colin Rafferty

Jeff Miller writes: > I believe I have found the culprit. If you look at xlwmenu.c, function > "string_width_u", you'll see that it has a 64 long char array called > newchars. I don't follow all that's going on in here, but this function > appear to clean up the text string and null-terminate it. > The 64 char arry obviously limits the length of the popup menu, so I'm > jacking it up to 256. It is always a bad idea to put limits on strings. > Seocond, there is an off-by-one count on the null termination of > newchars. So I fixed that too. Here is a patch that also works, but sets no limit on the length of the string. [ Note that I used text/plain instead of application/octet-stream]

1 participants
0 comments

Try to execute text in customize buffer 26 years, 5 months

David Bush

I am seeing something in a customize buffer that is similar to what Andy Piper reported last week. If I get into the cursor customization buffer from the menubar and select the value menu, then click on Bar Cusror ( 2pixels) I get in the minibuffer: Invalid function "Bar Cursor (2 pixels)" david

1 participants
0 comments

Re: Info warnings 26 years, 5 months

Michael Sperber

>>>>> "Oscar" == Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch> writes: >>>>> "Michael" == Michael Sperber <sperber(a)informatik.uni-tuebingen.de> writes: Oscar> Having dir files considerably speeds up the initial building of the Oscar> Info buffer as you don't need to open all the info files and look for Oscar> the ＠direntry in each of them. Michael> Sure, but the time it takes to rebuild is very acceptable out here. Michael> My info buffer has 307 lines from 11 separate info directories. Oscar> Well on my Sparc 5 with info files mounted through NFS the delay *is* Oscar> noticeable. I didn't say it's not noticeable, I said it's acceptable. Michael> So---can I force it to autogenerate without trying to write? Oscar> Currently no. But you won't see the difference, even if I implement it. With Oscar> (setq Info-rebuild-outdated-dir 'always), the existence or non-existence of dir Oscar> files is completely transparent to the user be the generated dir's written to Oscar> disk or not. You won't notice anything. Sure I notice something: It leaves `dir' files lying around where there shouldn't be any. XEmacs is not the only tool that groks the info format. -- Cheers =8-} Chipsy Friede, Völkerverständigung und überhaupt blabla

1 participants
0 comments

Re: Info warnings 26 years, 5 months

Oscar Figueiredo

>>>>> "Michael" == Michael Sperber <sperber(a)informatik.uni-tuebingen.de> writes: Oscar> Having dir files considerably speeds up the initial building of the Oscar> Info buffer as you don't need to open all the info files and look for Oscar> the ＠direntry in each of them. Michael> Sure, but the time it takes to rebuild is very acceptable out here. Michael> My info buffer has 307 lines from 11 separate info directories. Well on my Sparc 5 with info files mounted through NFS the delay *is* noticeable. Oscar> If you (setq Info-rebuild-outdated-dir 'always) then you can simply Oscar> ignore the existence of dir files, they're used as a cache and nothing more. Oscar> With this setting XEmacs never issues a warning about dir files, it writes them Oscar> to disk when it can and builds them as temporary when it can't. Michael> So---can I force it to autogenerate without trying to write? Currently no. But you won't see the difference, even if I implement it. With (setq Info-rebuild-outdated-dir 'always), the existence or non-existence of dir files is completely transparent to the user be the generated dir's written to disk or not. You won't notice anything. Oscar

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 June 1998