Re: [hylafax-users] Some faxes received at recipient blank

[Lee Howard <faxguy@xxxxxxxxxxxxxxxx>]

Daren Russell wrote:
> On 29/04/10 13:23, Aidan Van Dyk wrote:
>   
> >   
> >     
> > > >       +    -dMaxStripSize=0 \
> > > >       
> > > >         
> >   
> >     
> > > > This should work on both versions of GS, and make the new version of GS
> > > > write a single unlimtied strip that will work.  I thouls eliminate the
> > > > need for Use2D, or the tiffcp hack...
> > > >       
> > > >         
> >   
> >     
> > > Downgrading has worked for now, so it is currently off of my "priority"
> > > list!  When I have sorted out some other issues with a new server, I
> > > shall be going back to this and will give this a try when re-updating
> > > ghostscript (probably early next week).
> > >     
> > >       
> > OK, please do try it.  Just to give you more background.  Ghostscript
> > traditionally had that MaxStripSize set to 0.  It was changed in 8.71
> > (I think) to 8192.   And as soon as 8.71 was released, and they
> > realized they broke everything for everybody, they changed it.  They
> > didn't change it to 0 (unlimited), but to 1Million (or something big
> > like that).
> >
> > So, ghostscript 8.71 is affected.  Newer GS (already in fedora
> > repositories) bumps this up to a value large enough to hold a page if
> > tiff data.  And previous to 8.71 it was unlimitted.  So by next week,
> > your ports might have it fixed already and you not even need
> > -dMaxStripSize....
> >
> > But we want to make sure that -dMaxStripSize=0 works on "all"
> > ghostscripts, so if you could test that on your curren tone (and when
> > upgraded), I Still would appreciate it.
> >     

MaxStripSize is found in 'strings' output against Ghostscript 6.51, 
which is the oldest version I have available to me.  I think it's safe 
to assume that anyone running an updated HylaFAX is likely to have a 
compatible Ghostscript.

> This morning I updated ghostscript again to 8.71, and added the
> -dMaxStripSize option to pdf2fax and ps2fax.  We haven't used the fax
> much today, but the two that went this afternoon (which always seemed to
> fail before) seem to have gone through ok, so it appears this small
> change has fixed the issue with the change.

I think it's clear that default Ghostscript behavior has diverged from 
where HylaFAX has always expected it to be.  The -dMaxStripSize=0 option 
will apparently be necessary to restore the historic behavior... 
although I suspect that for *most* cases the 1MB default will be adequate.

All of this said, I concluded that it's not wise to forever expect 
Ghostcript output to be formatted as particularly as fax requires.  So 
while we can restore historic behavior by adding the -dMaxStripSize=0 
option I feel it's wise for HylaFAX to double-check the Ghostscript 
output anyway... and if it's not formatted as-needed to then correct it 
using libtiff tools or whatever.  HylaFAX already checks TIFFs in this 
fashion when they're submitted from the client as a TIFF... so this 
process of checking TIFF formatting is really nothing new to HylaFAX.  
That way, if Ghostscript ever does something else like this we don't 
have such a catastrophe on our hands.

Thanks,

Lee.


____________________ HylaFAX(tm) Users Mailing List _______________________
 To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
On UNIX: mail -s unsubscribe hylafax-users-request@xxxxxxxxxxx < /dev/null
 *To learn about commercial HylaFAX(tm) support, mail sales@xxxxxxxxx*