Re: [hylafax-users] AW: Problems receiving facsimiles, modems don't return to idle state

[Lee Howard <faxguy@xxxxxxxxxxxxxxxx>]

On 2003.11.20 11:58 Thorsten Büker wrote:

> > The default ModemPageDoneTimeout is a rather large integer (180000).
>
> > So it is bigger than 16-bits, so it requires a 32-bit
> > allocation by the
> > compiler. It and ModemPageStartTimeout are the only default timeout
> > values that large.  You could try setting it lower, say at 60000
> (one
> > minute, and within a 16-bit range) to see if the size of the timeout
>
> > value is a factor.
>
> To check if the fault occurs more often in connection to a quicker
> timeout?

No.  To check to see if it a problem with the default value exceeding 
the allocated space that was given it by the compiler.  If it never 
occurs with the 60000 setting then it would indicate a typing problem 
within the HylaFAX code.

> > I can send you code patches to add some debugging info in
> > there so that
> > we can see what's actually going on.  If you can make use of them
> and
> > want to try it out, let me know.
>
> Yes, this would be nice.

Please try the attached patch.  It will add a fair amount of logging.  
This is expected and necessary, unfortunately, in order to debug this.  
When you get your next lockup please send the associated session log.

> Will the patches also work on the sources shipped by SuSE?

The patch is based on CVS-HEAD (future 4.2.0), but should apply to 
SuSE's 4.1.5 and probably much earlier.

> > The catalyst to the problem, however, appears to be the lack of an
> > expected modem response.  In Class 2.0 you'll need to consult
> > the modem
> > manufacturer for a firmware fix.  So HylaFAX can fix the "stuck"
> > faxgetty, but the modem manufacturer must fix the modem error.
>
> Actually the modems (ELSA Microlink 28.8 TQV) worked fine under W2k.
> Unfortunately ELSA meanwhile became insolvent, I don't know about
> available firmware fixes.

I highly suspect that the modems did the same thing with Windows 2000; 
only the difference being that the fax program you used handled the 
modem problem with greater ease (like, not causing the program to hang).

Lee.

Attachment: debug.patch
Description: Binary data