Re: BUG REPORT tin 1.7.2 release 20030929 ("Berneray") [UNIX]

On Wed, Oct 01, 2003 at 11:36:40AM +0200, Sebastian Bork wrote:
> CFG3 : debug=0, threading=2
> 
> Please enter _detailed_ bug report, gripe or comment:
> When you try to add a killfile entry with <Ctrl>+<k> and then select the
> Message-ID to be matched (I tried with "last"), the entry that shows up
> in the killfile contains the first Message-ID of the thread instead of
> the Message-ID of the article that was displayed when you hit <Ctrl>+<k>
> (although the correct Message-ID is displayed in the dialog).

when I tried to reproduce this I couldn't till I switched threading
from both to refereces only.

> I think this might be a side-effect of the patch to the quick-kill and
> quick-select feature, but that's only a wild guess.

on a first quick look I'd say it's caused by add_filter_rule() which
is used for both quick-kill and menu-kill without having a flag to differ
between both (to skip the quick-kill logic added a while ago in the
menu case). anyone more experienced with the filter stuff willing to
look at this?

urs
--

-- 
"Only whimps use tape backup: _real_ men just upload their important stuff
 on ftp, and let the rest of the world mirror it ;)" - Linus

[tin 1.7.x] new snaps

I've made new snaps including Michaels and Jasons latest patches:
<ftp://ftp.tin.org/pub/news/clients/tin/unstable/snapshots/≥

critical bugs (as the stable branch is also affected and they might
lead to article loss) not yet on the TODO-list

- article loss when resizing the final-post screen
  see <nntp://news.tin.org/20030927141249.GA15157 <at> akk10.akk.uni-karlsruhe.de>

- generating wrong filter entry via menu
  see <nntp://news.tin.org/200310010936.h919ae1r006383 <at> eris.sebi.org>

urs
--

-- 
"Only whimps use tape backup: _real_ men just upload their important stuff
 on ftp, and let the rest of the world mirror it ;)" - Linus

Re: [tin 1.7.x] new snaps

Urs Janßen schrieb:
> - article loss when resizing the final-post screen
>   see <nntp://news.tin.org/20030927141249.GA15157 <at> akk10.akk.uni-karlsruhe.de>

The article is _not_ lost (unless you press 'q') -- just checked
this to be sure. It's only that the posting menu is overwritten with
the article/thread/group/selection page when you resize the window.

See my previous mail about this topic. You "only" have to introduce
another signal_context value and handle this appropriatly.

(fwd) Modified url_handler.sh script

----- Forwarded message from David Highley <dhighley <at> highley-recommended.com> -----

I made a few modifications to the url_handler.sh script. I found that
the tr command was not translating upper case to lower case in the
method assignment line when run on a Solaris 9 system. The modification
I did has been tested on Solaris 9 and RedHat 9.

Then I added multiple possible browser and other programs for each
method. I know that there is no way to tell which preference order each
site would like but they should now be able to reorder the lines to fit
there site.

Finally if the netscape browser is older than version 6 it will fail if
the browser is not all ready running.

--

-- 

Regards,

David Highley		      Phone: (206) 669-0081
Highley Recommended, Inc.	FAX: (253) 838-8509
2927 SW 339th Street	      Email: dhighley <at> highley-recommended.com
Federal Way, WA 98023-7732	WEB: http://www.highley-recommended.com

Content-Description: executable shell script
#! /bin/sh
#
# example of how to call the appropriate viewer
# based on a script by Michael Elkins <me <at> cs.hmc.edu>
# 2001-01-31 <urs <at> tin.org>

[tin 1.7.x][PATCH] Make page footer multibyte safe

Hello,

this patch makes the page footer multibyte safe and removes one
occurance of MORE_POS.

Michael

 po/de.po   |    2 +-
 src/misc.c |   14 +++++++++++++-
 src/page.c |   20 ++++++++++++++++++--
 3 files changed, 32 insertions(+), 4 deletions(-)
diff -Nurp tin-1.7.2/po/de.po tin-1.7.2.o/po/de.po
--- tin-1.7.2/po/de.po	2003-10-02 11:53:03.000000000 +0200
+++ tin-1.7.2.o/po/de.po	2003-10-05 01:15:45.000000000 +0200
 <at>  <at>  -2174,7 +2174,7  <at>  <at>  msgstr	"newsrc-Datei erfolgreich geschri

 #: src/lang.c:516
 msgid	"-- Next response --"
-msgstr	"- N�chste Antwort -"
+msgstr	"-- N�chste Antwort --"

 #: src/lang.c:517
 #, c-format
Files tin-1.7.2/src/.filter.c.swp and tin-1.7.2.o/src/.filter.c.swp differ
diff -Nurp tin-1.7.2/src/misc.c tin-1.7.2.o/src/misc.c
--- tin-1.7.2/src/misc.c	2003-09-25 19:59:19.000000000 +0200
+++ tin-1.7.2.o/src/misc.c	2003-10-05 01:00:15.000000000 +0200
 <at>  <at>  -816,6 +816,10  <at>  <at>  draw_percent_mark(
 	long max_num)

(fwd) Bug#214405: tin: Segfault in which_thread

I can't reproduce that - anyone willing to have a look at this?

----- Forwarded message from Vincent Lefevre <vincent <at> vinc17.org> -----
To: Debian Bug Tracking System <submit <at> bugs.debian.org>

Package: tin
Version: 1:1.7.1+20030906-1
Severity: important

I sometimes get crashes with the current tin version, and I could
reproduce one. So, I recompiled tin from source, did a gdb ./src/tin
and here's what I get:

ay:~/tin-1.7.1+20030906> gdb ./src/tin
GNU gdb 5.3-debian
Copyright 2002 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "powerpc-linux"...
(gdb) run -ar

Now, I go to fr.comp.os.unix, type 'r' (as all the articles are read),
move the cursor here:

->  32       2  Un soi-disant "spécialiste" ne c onnait pa  Youri

and hit the right arrow key. -> Segmentation fault:

Re: (fwd) Bug#214405: tin: Segfault in which_thread

On Mon 06 Oct 2003 13:46:29 GMT, Urs Janßen wrote:
> I can't reproduce that - anyone willing to have a look at this?

I can't reproduce it either, but I see the problem:
The problem lies within group.c:group_right() and tinrc.auto_list_thread
turned off. 

> ----- Forwarded message from Vincent Lefevre <vincent <at> vinc17.org> -----
> Now, I go to fr.comp.os.unix, type 'r' (as all the articles are read),
> move the cursor here:
> 
> ->  32       2  Un soi-disant "spécialiste" ne c onnait pa  Youri
> 
> and hit the right arrow key. -> Segmentation fault:
> 
> Program received signal SIGSEGV, Segmentation fault.
> 0x10056600 in which_thread (n=-1) at thread.c:942
> 942             for (i = n; arts[i].prev >= 0; i = arts[i].prev)
> (gdb) where
> #0  0x10056600 in which_thread (n=-1) at thread.c:942
                                  ^^^^
> #1  0x1001c968 in group_right () at group.c:131

next_unread() (group.c:129) returns -1 and which_thread get called with
this value (group.c:131). which_thread tries to access arts[-1] which
causes a segmentation fault on powerpc but apparently not on i386.

A fix is to change the order within the if statement in group.c:131 from
| if (grpmenu.curr == which_thread(n) && n >= 0)
to

[tin 1.6.x, 1.7.x] patch: word_highlight change in the Menu

the patch below (made against my 1.7.2-snapshot, but should also work
with 1.6.2) fixes not propagating back tinrc.word_highlight changes
at runtime.

--- config.c.o	Wed Oct  8 01:00:59 2003
+++ config.c	Wed Oct  8 01:04:23 2003
 <at>  <at>  -1893,6 +1893,11  <at>  <at> 
 							break;
 #endif /* XFACE_ABLE */

+						/* word_highlight */
+						case OPT_WORD_HIGHLIGHT:
+							word_highlight = tinrc.word_highlight;
+							break;
+
 						/*
 						 * the following are boolean and do not need further action (if I'm right)
 						 *
 <at>  <at>  -1932,7 +1937,6  <at>  <at> 
 						 * case OPT_USE_KEYPAD:
 #endif
 						 * case OPT_USE_MOUSE:
-						 * case OPT_WORD_HIGHLIGHT_TINRC:
 						 * case OPT_WRAP_ON_NEXT_UNREAD:
 						 */

urs
--

-- 
"Only whimps use tape backup: _real_ men just upload their important stuff
 on ftp, and let the rest of the world mirror it ;)" - Linus

Re: [tin 1.6.x, 1.7.x] patch: word_highlight change in the Menu

> comments? which special-key options should be removed? or should the
> just all propagate back to the tinrc? (don't use two diffrent
> state-vars).

I think we had this debate some years ago, and the answer was "it depends"
Some of the settings are very obviously transient, eg show_only_unread
and you would not want them to 'stick'. Others are more debateable.

There is a fair bit of cleanup to do in tinrc - there is the TODO for
propogation of changes to tinrc back to global attributes for example.
Only a few vars are handled properly for this case as of now.

Jason.

[tin 1.7.x][PATCH] Some code cleanup

Changes:
* fix a bug in strunc() in !MULTIBYTE-ABLE-case: add "..." only when the
  string was truncated
* build_[st]line: allocate 'buffer'
* use global t_group *curr_group instead of CURR_GROUP
  (done for page.c, group.c and thread.c)

Michael

 group.c  |   68 +++++++++++++++++++------------------
 page.c   |    2 -
 string.c |    6 ++-
 thread.c |  114 ++++++++++++++++++++++++++++++++-------------------------------
 4 files changed, 100 insertions(+), 90 deletions(-)
diff -Nurp tin-1.7.2/src/group.c tin-1.7.2.o/src/group.c
--- tin-1.7.2/src/group.c	2003-10-06 15:37:18.000000000 +0200
+++ tin-1.7.2.o/src/group.c	2003-10-09 14:16:10.000000000 +0200
 <at>  <at>  -986,13 +986,13  <at>  <at>  toggle_read_unread(
 	int n, i = -1;

 	if (force)
-		CURR_GROUP.attribute->show_only_unread = TRUE;	/* Yes - really, we change it in a bit */
+		curr_group->attribute->show_only_unread = TRUE;	/* Yes - really, we change it in a bit */

 	wait_message(0, _(txt_reading_arts),
-		(CURR_GROUP.attribute->show_only_unread) ? _(txt_all) : _(txt_unread));
+		(curr_group->attribute->show_only_unread) ? _(txt_all) : _(txt_unread));

 	if (grpmenu.curr >= 0) {