Urs Janßen | 7 May 17:46 2009

Re: 1.9.4 does not compile outside source directory

On Thu, May 07, 2009 at 12:21:27PM +0200, Volker Borchert wrote:
> 
> pcre/Makefile.in has two deficiencies with respect to compilation
> outside the source tree:
> 1. It lacks a -I$(srcdir)
> 2. It does not handle pcre_chartables.c which is generated in the
>    build tree specially
> 
> The following works for me:
> 
> ------- Makefile.in -------
> *** -   Thu May  7 11:53:18 2009
> --- pcre/Makefile.in    Tue May  5 20:08:17 2009
> ***************
> *** 31,37 ****
>         -DMATCH_LIMIT_RECURSION=10000000 \
>         -DNEWLINE=10 \
>         -DPOSIX_MALLOC_THRESHOLD=10 \
> !        <at> DEFS <at>  -I. -I../include
>   
>   RANLIB        =  <at> RANLIB <at> 
>   RM    = rm
> --- 31,37 ----
>         -DMATCH_LIMIT_RECURSION=10000000 \
>         -DNEWLINE=10 \
>         -DPOSIX_MALLOC_THRESHOLD=10 \
> !        <at> DEFS <at>  -I. -I../include -I$(srcdir)
>   
>   RANLIB        =  <at> RANLIB <at> 
>   RM    = rm
(Continue reading)

Urs Janßen | 7 May 19:37 2009

Re: BUG REPORT tin 1.9.4 release 20090211 ("Rieclachan") [UNIX]

On Thu, May 07, 2009 at 02:24:02PM +0200, Volker Borchert wrote:
> Version: tin 1.9.4 release 20090211 ("Rieclachan") May  7 2009 13:15:00
> Platform:
>         OS-Name  = "sunos4.1.4"
> Any editor invocation is erroneously classified as "failed" in invoke_cmd().
> 
> Neither USE_SYSTEM_STATUS nor IGNORE_SYSTEM_STATUS are defined:
> ./sparc-sun-sunos4.1.4/include/autoconf.h:/* #  undef USE_SYSTEM_STATUS */
> ./sparc-sun-sunos4.1.4/include/autoconf.h:/* #  undef IGNORE_SYSTEM_STATUS */
> 
> tin-1.5.5 works fine on the same system with the same shell and editor.

IGNORE_SYSTEM_STATUS is set when --enable-broken-system-fix is given at
configure time to override broken system()-status implementations. the name
of USE_SYSTEM_STATUS might be a bit confusing as IIRC if it's defined the
system()-status is not evaluated and if it's not defined it is. so in your
case configure detects a usable system()-status (USE_SYSTEM_STATUS not
defined) but it doesn't work and you should configure with
--enable-broken-system-fix like on m88k-dg-dgux5.4R3.00, mips-sei-seiux3.2,
mips-mips-riscos4sysv, m68k-next-openstep4, m68k-apple-aux3.1.1,
m68k-next-nextstep3, m68k-sony-newsos, mips-sony-bsd, hppa1.0-hp-hpux8.00
and m68k-sun-sunos4.1.1_U1 (see http://www.tin.org/builds.html).

we should fix the CF_FUNC_SYSTEM configure macro so the additional configure
option is not needed on those old OSs.

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
(Continue reading)

Urs Janßen | 15 May 14:31 2009

[PATCH] tin 1.9.x threading problem

the patch below fixes a crash with references/both-threading in the newsgroup
alue.devel available on flowerpot.kaijanaho.info (in the thread starting
with Message-Id: <3319583.LvFx2qVVIh <at> valdyas.org> / Subject: Feature
request: get TLS and authorisation working). I'm not sure if it's the
right approach, maybe it should be fixed somewhere else - a review would be
nice.

=== modified file 'src/refs.c'
--- src/refs.c	2008-12-30 19:36:57 +0000
+++ src/refs.c	2009-05-15 11:14:03 +0000
 <at>  <at>  -56,7 +56,7  <at>  <at> 
 static struct t_msgid *add_msgid(int key, const char *msgid, struct t_msgid *newparent);
 static struct t_msgid *find_next(struct t_msgid *ptr);
 static struct t_msgid *parse_references(char *r);
-static t_bool valid_msgid(const char *msgid);
+static t_bool valid_msgid(char *msgid);
 static unsigned int hash_msgid(const char *key);
 static void add_to_parent(struct t_msgid *ptr);
 static void build_thread(struct t_msgid *ptr);
 <at>  <at>  -176,15 +176,16  <at>  <at> 
  * Checks if Message-ID has valid format
  * Returns TRUE if it does, FALSE if it does not
  *
- * TODO: combine with post.c:damaged_id()?
+ * TODO: combine with post.c:damaged_id()
  */
 static t_bool
 valid_msgid(
-	const char *msgid)
+	char *msgid)
(Continue reading)

Dennis Preiser | 21 May 15:05 2009
Picon
Picon

[PATCH] cancel_article() corrections and minor other changes

Hello,

attached a (small) patch which deals with the following issues:

post.c:cancel_article():

- When canceling an own article and FORGERY + EVIL_INSIDE are defined  
no message-id is generated.
- If FORGERY is not defined, redraw_screen never becomes TRUE. Thus,  
tin doesn't switch back to article level after canceling an article  
or after 'q'uitting the cancel process (draw_page() isn't called in  
page.c:show_page():755 because cancel_article() returns FALSE).

Other changes:

- fixed a typo in auth.c:authinfo_plain()
- removed an unnecessary variable in attrib.c:add_scope() (only used  
to return zero)
- insert a clear_message() in newsrc.c:read_newsrc() in order to  
clear the line proper after 'reread active file' ('Y')

Dennis

Attachment (20090521.diff.gz): application/x-gzip, 1314 bytes

Gmane