.... don't forget to set the option for 'configure':
--with-curses-lib=/

It defaults to '/usr' instead of '/', so some library won't be found, so
compilation is stopped with errors. At least this holds for slackware-10.0
based environment.

Still, news works best for me with 'suck' and 's-news'. There is the minor
but a bit nasty trade-off, so 'tin' won't show one single new posting, or
the first new posting, in a newsgroup (after 'lmove'-part of 'suck' has put
it into the spool and has incremented the counters of '.newsrc').

Here the old crux of programming seems to re-appear, that an event at offset
_zero_ should be the _first_ event but isn't recognized as such. With one
new posting, 'tin' switches to the overview of a group but won't set a mark
for an unread posting. Only with two new postings, one mark is set. It might
be guessed, that this accomodates to the behaviour of 'inn' as a server?

Maybe someone else has found another trick for bypassing (or I've overseen
something)? I use a little script, so if there is new postings (the counters
in the file 'active' differ), so the counter in '.newsrc' is modified. It's
not really nice as a solution and does not work for all instances, but
somehow it catches-up a bit, so:

-----------------------------------------

#!/bin/sh
na="/var/lib/news/active"
nb="/usr/lib/news/.newsrc"
nc="/usr/lib/news/.newsrc.n"

if [ -f $na ] ; then {
if [ -f $nb ] ; then {

while read z ; do {
zi=`echo $z | cut -d" " -f1 |cut -d: -f1` ;
while read y ; do {
yi=`echo $y | cut -d" " -f1 |cut -d: -f1` ;

if [[ $zi == $yi ]]; then {
let zn=`echo $z | cut -d" " -f2` ;
let yn=`echo $y | cut -d" " -f2 |cut -d- -f2 |cut -d, -f1 `;

if [ $yn -lt $zn ]; then {
# echo -n $yn " "
echo $yi
let yx=$yn-1 ;
let yn=$yx
} fi ;
echo $yi": 1-"$yn >>$nc ;

} fi ;
} done < $nb ;
} done < $na ;

mv -f $nb "$nb.o"
mv $nc $nb
} fi ;
} fi ;

-----------------------------------------

Ewald Pfau <anderswo@gmx.net> wrote:
> .... don't forget to set the option for 'configure':
> --with-curses-lib=/
>
> It defaults to '/usr' instead of '/', so some library won't be found, so
> compilation is stopped with errors. At least this holds for slackware-10.0
> based environment.
>
Compiled tin-1.7.8 in Slack 10.0, without the above option, with no errors.

-- Julio

Julio Uehara <julueh@tma-1.discovery.net>:
> Ewald Pfau <anderswo@gmx.net> wrote:
>> .... don't forget to set the option for 'configure':
>> --with-curses-lib=/
>
> Compiled tin-1.7.8 in Slack 10.0, without the above option, with no errors.

Maybe it is related to the kernel version (I see 2.6.11 running)?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ewald Pfau <anderswo@gmx.net> wrote:
> Julio Uehara <julueh@tma-1.discovery.net>:
>> Ewald Pfau <anderswo@gmx.net> wrote:
>>> .... don't forget to set the option for 'configure':
>>> --with-curses-lib=/
>>
>> Compiled tin-1.7.8 in Slack 10.0, without the above option, with no errors.
>
> Maybe it is related to the kernel version (I see 2.6.11 running)?

It shouldn't. I've compiled it under 2.4.30 and 2.6.11.6 also
without any problems, and that is with --with-screen=ncurses (which you
should be using instead of --with-curses-dir).

Could you post the errors you're getting?

BL.
- --
Brad Littlejohn | Email: tyketto@sbcglobal.net
Unix Systems Administrator, | tyketto@ozemail.com.au
Web + NewsMaster, BOFH.. Smeghead! | http://www.sbcglobal.net/~tyketto
PGP: 1024D/E319F0BF 6980 AAD6 7329 E9E6 D569 F620 C819 199A E319 F0BF

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFCUaiMyBkZmuMZ8L8RAqKVAJ41ufISNVGmpC7Bsn3pyu w2qmX/WgCgy2fZ
QjCq/XpF1FHxZZLdKd9ItoY=
=r/UZ
-----END PGP SIGNATURE-----

A Guy Called Tyketto <tyketto@sbcglobal.net.invalid>:

> -----BEGIN PGP SIGNED MESSAGE-----
gone

> It shouldn't. I've compiled it under 2.4.30 and 2.6.11.6 also
> without any problems, and that is with --with-screen=ncurses (which you
> should be using instead of --with-curses-dir).
>
> Could you post the errors you're getting?

So I just tried to make'em up once more, to verify. And now I am stuck: it
just compiles. Sorry, I don't know why, so I have to stand corrected: it
compiles as it should.

In the first run, it had cost me half an afternoon, applying little changes
in 'configure.in' and finally in 'conf-tin', where I ended with the proposed
switch, where to look for ncurses-lib. The errors I got all the time have
been unresolved library calls to ncurses (mainly 'tputs'), so it was
reasonable that this switch solved that problem.