Re: slrn 1.0 [2007-08-25]?

Thomas Wiegner wrote:
> * Martin Trautmann (traut <at> gmx.de) wrote [Sun, 070729 16:21]:
> 
>> No, I wrote above: "./configure" ... as it is claimed within 
>> doc/INSTALL.unix
> 
>> autogen.sh shows the same error, looking for an install.sh or install-sh
> 
> The 'automake --foreign --add-missing' normally copies the
> install.sh from /usr/share/automake-XX to the autoconf directory, if
> this does not work, something with your automake installation is
> messed. You should check if install-sh is somewhere installed on you 
> computer, it should be located somewhere like:
> 
> | /usr/share/automake-1.9/install-sh
> | /usr/share/automake-1.10/install-sh

It's in there, but it did not copy it to autoconf.

> If you have this files installed and it does not work for some
> reason, you could also try to take the debian package with the
> broken links and replace the links with the according files of your
> automake installation.

I might try the debian package - but we are talking about the release of 
a 1.0 version here. Am I wrong to expect that this should find its way 
on its own?

Current errors:

Re: automake 1.10 and autoconf 2.61 are required for cvs compilation

On Tuesday 31 July 2007 17:05, Thomas Wiegner wrote:

> I was able to compile the cvs slrn with these versions of autoconf
> and automake:

> | autoconf (GNU Autoconf) 2.59
> | automake (GNU automake) 1.9

But the problems other people have when compiling slrn from CVS will disappear 
if there's a .pl2 or 1.0-rc1 release, preferably with your charset patch 
included.

The only reason non-developers like me are using CVS code is that there are 
fixes available in CVS for .pl1 problems.

Meanwhile, I've been studying the docs. There are a few errors in manual.txt 
and $SLRNHELP/help.txt that need to be fixed; and there are also errors in 
the coresponding pages at www.slrn.org. I hope to have at least a list of 
errors, or (better still) a patch for the SGML source ready within a few 
days.

For instance, the instructions for resizing the article window with ^ and 
Ctrl-^ seem to be the wrong way round.

If Thomas (Schultz, not Wiegner) isn't too busy, I'd really like to see a .pl2 
release soon... and maybe a 1.0 before the end of 2007?

--

-- 
PJR 

Re: automake 1.10 and autoconf 2.61 are required for cvs compilation

* Peter J Ross (pjr.spam <at> gmail.com) wrote [Wed, 070801 19:45]:

> But the problems other people have when compiling slrn from CVS
> will disappear if there's a .pl2 or 1.0-rc1 release, preferably
> with your charset patch included.

> The only reason non-developers like me are using CVS code is that
> there are fixes available in CVS for .pl1 problems.

To ease the pain a little bit, I put a tarball of the .pl2 CVS
version on the foory website. The tarball was created with "make
dist", so there should be no problems with all the autoconf and
automake issues.

To make "make dist" work I had to replace the line "src/chmap.c"
with "src/charset.c" in po/POTFILES.in. I think this is the
reason, why the debian tarball of .pl2 is broken as "make dist"
did not work. 

> If Thomas (Schultz, not Wiegner) isn't too busy, I'd really
> like to see a .pl2 release soon... and maybe a 1.0 before the
> end of 2007?

This would be really nice.

 Thomas

--

-- 
slrn charset patches: http://www.foory.de/thw/slrn/

UTF-8-problems; more tolerance for invalid charsets

Hi all,

thanks to some excellent help I finally managed to compile slrn for linux and Mac.

Now I feel that some utf problems where fixed, while others are even worse than before

- subject lines are cut at 8bit chars:

  Some subjects just got truncated at the first 8bit char (linux, S-Lang 1)

- text is cut at 8bit chars:

  Sometimes it's even the end of each text line which is cut (linux, S-Lang 1)

- non-ASCII is written as <XY> hex codes

  ... which is even worse since this may corrupt columns widths
  This may be somehow related to S-Lang? On the Mac I got
  S-Lang Library Version: 2.1.1 which causes

  1  [Ale Ölerss] Subject
  2  [B<FC>rbel Blu] Other thread

  On linux I got the old S-Lang Library Version: 1.4.9
  Which results in
  1  [Ale Ölerss] Subject
  2  [Bärbel Bl] Other thread

- other S-Lang fragments

Re: UTF-8-problems; more tolerance for invalid charsets

* Martin Trautmann (traut <at> gmx.net) wrote [Mon, 070806 16:13]:

> thanks to some excellent help I finally managed to compile slrn
> for linux and Mac.

> Now I feel that some utf problems where fixed, while others are
> even worse than before

[ lot of problems snipped ]

First of all, you'll run into problems reading postings with missing
charset declarations containing umlauts, make sure you have at least
my fallback_charset patch installed.

As you are using a Mac, do you have libiconv installed? If I
remember correctly, the iconv support in OSX is not included in
libc like it is in Linux. If your slrn is not compiled against
libiconv I don't think it will work very good with *any* charset
other than *us-ascii*.

Are you running slrn in a UTF-8 terminal? slrn detects the
termencoding from the locale settings? The locale settings can be
overridden by the undocumented config option "charset Display".
Set this to "iso-8859-15" or "utf-8" depending on you terminal
capabilities.

 Cheers,

  Thomas

Re: UTF-8-problems; more tolerance for invalid charsets

> First of all, you'll run into problems reading postings with missing
> charset declarations containing umlauts, make sure you have at least
> my fallback_charset patch installed.

Thanks, the header view is much more readable now, while the messages still look the same

> As you are using a Mac, do you have libiconv installed? 

I'll check later on. I've seen ./configure options here, but no hint in the doc what to use.

> If I
> remember correctly, the iconv support in OSX is not included in
> libc like it is in Linux. If your slrn is not compiled against
> libiconv I don't think it will work very good with *any* charset
> other than *us-ascii*.

Since it does work very well in properly defined charsets, I guessed that this behaviour would be ok. I see
the same behaviour on linux where you expect that it would be included by default.

> Are you running slrn in a UTF-8 terminal? 

Yes.

> slrn detects the termencoding from the locale settings? 

I hope so:  LANG=en_US.UTF-8
etc. for all the LC_*, but LC_ALL is undefined.

> The locale settings can be
> overridden by the undocumented config option "charset Display".

slrn hangs on X-Face header

Hello,

I'm using slrn 0.9.8.1pl2.1 [2007-07-17] (Debian package build with
Thomas' script "build_deb_from_cvs.sh").  I encountered several
postings that let slrn eat all memory including swap until killed by
the kernel.  I noticed that all these postings have in common that
they contain an "X-Face:" header and when I remove this header in my
local news spool slrn can display these postings just fine.

An example for such a header:

X-Face: /#02M#7?aBv8;\1p(MfZIgg5-R^.<T%UtJ6 <at> tef|[6ey^X<u <at> ^;8.e$<=?utf-8?q?i*fEQyPiaahnMV=0A=09saSMj2Jd=27=23-!g=25Q=279=3BSE=5Ee/X=60VgPp?=
 =?utf-8?q?T=7DgUNvr/=5F=7Be?=>mU?}d9NJDx <at> L/~j <at> mP~s=hQ?,
 =?utf-8?q?=24=60=26c=0A=09=7EU?= <at> 2u%yj.N+&g!M)F!l~Z`Yp.UX(VdMsy97G"lJe4xmg-LgwedJMSCp{^u?YdC<
 =?utf-8?q?54=5BjP=27=3A=0A=09fO?=

as found in <news:8NrAE-1CA-23 <at> gated-at.bofh.it>

Is this a bug or a misconfiguration on my part?

Greetings,
Martin

--

-- 
pub   1024D/DCF157DF 2002-07-21
      Key fingerprint = A2D4 DDCC 2426 93F2 63C2  AEED C4EE 944E DCF1 57DF

Today is Pungenday, 72. day of Confusion in the YOLD of 3173

Re: slrn hangs on X-Face header

* Martin Grandrath (lists <at> grandrath-net.de) wrote [Mon, 070806 21:23]:

> [ slrn crashes with x-face ]

> An example for such a header:

> X-Face: /#02M#7?aBv8;\1p(MfZIgg5-R^.<T%UtJ6 <at> tef|[6ey^X<u <at> ^;8.e$<=?utf-8?q?i*fEQyPiaahnMV=0A=09saSMj2Jd=27=23-!g=25Q=279=3BSE=5Ee/X=60VgPp?=
>  =?utf-8?q?T=7DgUNvr/=5F=7Be?=>mU?}d9NJDx <at> L/~j <at> mP~s=hQ?,
>  =?utf-8?q?=24=60=26c=0A=09=7EU?= <at> 2u%yj.N+&g!M)F!l~Z`Yp.UX(VdMsy97G"lJe4xmg-LgwedJMSCp{^u?YdC<
>  =?utf-8?q?54=5BjP=27=3A=0A=09fO?=

I'm not too deep into the x-face standard, but I must say, this
looks not like a correct X-Face header. As far as I understand it,
there should be only us-ascii headers, so this header is corrupt.

For me slrn does not freeze, but I don't have the display of X-faces
headers turned on.

When I take a look at the headers from within slrn it shows me this
line as

| X-Face: /#02M#7?aBv8;\1p(MfZIgg5-R^.<T%UtJ6 <at> tef|[6ey^X<u <at> ^;8.e$<i*fEQyPiaahnMV

but when I'm piping the article to 'less' it looks like you wrote.

Do you use slrnface? Does slrn also crash when not using slrnface? I
think the problem might be slrnface.

Cheers,

Re: slrn hangs on X-Face header

On Monday 06 August 2007 20:23, Martin Grandrath wrote:
> Hello,

Hello. I have no solution, merely some preliminary diagnostics to suggest.

> I'm using slrn 0.9.8.1pl2.1 [2007-07-17] (Debian package build with
> Thomas' script "build_deb_from_cvs.sh").  I encountered several
> postings that let slrn eat all memory including swap until killed by
> the kernel.  I noticed that all these postings have in common that
> they contain an "X-Face:" header and when I remove this header in my
> local news spool slrn can display these postings just fine.

I've experienced no problems with posts containing an X-Face: header. I'm 
using TW's patch too, but I'm compiling from CVS source, not using a .deb.

> An example for such a header:
>
> X-Face:
> /#02M#7?aBv8;\1p(MfZIgg5-R^.<T%UtJ6 <at> tef|[6ey^X<u <at> ^;8.e$<=?utf-8?q?i*fEQyPia
>ahnMV=0A=09saSMj2Jd=27=23-!g=25Q=279=3BSE=5Ee/X=60VgPp?=
> =?utf-8?q?T=7DgUNvr/=5F=7Be?=>mU?}d9NJDx <at> L/~j <at> mP~s=hQ?,
> 
> =?utf-8?q?=24=60=26c=0A=09=7EU?= <at> 2u%yj.N+&g!M)F!l~Z`Yp.UX(VdMsy97G"lJe4xmg-
>LgwedJMSCp{^u?YdC< =?utf-8?q?54=5BjP=27=3A=0A=09fO?=
>
> as found in <news:8NrAE-1CA-23 <at> gated-at.bofh.it>

I can't find that Message-Id, and slrn truncates the header when I try to post 
it to misc.test. (This is possibly because X-Face headers need extra escaping 
in order to be sent by slrn.)

Re: slrn hangs on X-Face header

Hello,

On Mon, Aug 06, 2007 at 10:02:36PM +0200, Thomas Wiegner wrote:
> I'm not too deep into the x-face standard, but I must say, this
> looks not like a correct X-Face header. As far as I understand it,
> there should be only us-ascii headers, so this header is corrupt.

Yes, that's what I thought.  But slrn shouldn't hang on bad input
though.

> Do you use slrnface? Does slrn also crash when not using slrnface? I
> think the problem might be slrnface.

I tried to use slrnface, but it didn't show anything at all, so I
removed it again, but the problem remained.

Gruß,
Martin

--

-- 
pub   1024D/DCF157DF 2002-07-21
      Key fingerprint = A2D4 DDCC 2426 93F2 63C2  AEED C4EE 944E DCF1 57DF

Today is Boomtime, 73. day of Bureaucracy in the YOLD of 3173

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.