flexfax: Re: [hylafax-devel] Re: Endless RTN -- seems to be fixed now!

[Dmitry Bely <dbely@mail.ru>]

Robert Colquhoun <rjc@trump.net.au> writes:

> >Folks, aside from "Bernd Proissl" <news@proissl.de>, has anyone checked
> >Dmitry's work?? If his analysis is correct, this is FANTASTIC! work, and may
> >well resolve one of the longest running issues with HylaFAX . . . the dreaded
> >RTN frequency.
> 
> Yes i have looked at it - the problem is the patch assumes that each 
> tiff/g3 page has the RTC characters appended to it which is not necessarily 
> true.

For Ghostscript it's always true, if there is no chopping (this check is
made in my patch)

>  Each page should be checked

I wrote it in the comment -- it's still experimental.

> by looking  at the last few characters 
> to see whether the RTC chars are present.

Wrong. RTC may not be byte-aligned (if we talk about general case of G3
TIFF), so you should decode the whole page to find it out.

All this does not really matter now. The fix can be made different ways,
e.g.

[---cut---]
--- util/ps2fax.gs.sh.orig	Tue Oct 13 00:47:48 1998
+++ util/ps2fax.gs.sh	Sun Mar 12 13:36:56 2000
@@ -59,6 +59,7 @@
 . etc/setup.cache
 
 PS=$GSRIP
+TIFFCP=$TIFFBIN/tiffcp	# part of the TIFF distribution
 
 fil=
 out=ps.fax		# default output filename
@@ -107,11 +108,16 @@
 # NB: unfortunately it appears to break valid PostScript;
 #     so it's been disabled.
 #
+tmp=$out.tmp
 $CAT $fil | $PS -q \
     -sDEVICE=$device \
     -dNOPAUSE \
     -dSAFER=true \
     -sPAPERSIZE=$paper \
     -r204x$vres \
-    "-sOutputFile=$out" \
-    -
+    "-sOutputFile=$tmp" \
+    - \
+&& $TIFFCP $tmp $out && rm -f $tmp
+#
+# tiffg3/tiffg32d bug: remove trailing RTC
+#
[---cut---]

The real question is: does this fix the RTN bug for those who suffers from
it (except me)? Personally I cannot test all fax machines in the world
without help of other Hylafax hackers. If it does, I'll make extremely
correct patch to satisfy you :-)

> PS I am glad someone else is looking at this i have had this problem on my 
> todo list since the end of 
> 1998(http://www.hylafax.org/archive/1998-12/msg00015.html) but have not had 
> any time to do anything about it :-(

Believe me, I've found it independently :-) But having known about such
obvious bug, do nothing for more then a year -- I'm very impressed, not to
say more ...

Hope to hear from you soon,
Dmitry