flexfax: PLEASE: fax to a fax-number, another one answers

[Giulio Orsero <giulioo@tiscalinet.it>]

***
I already posted about this a while  ago and I was told it cannot be a
hylafax problem, but it happened again and with another modem.
***

Linux kernel 2.0.33, hylafax 4.0pl1 (from the libc5 rpm hylafax-4.0-6).
_Please_ dont't simply tell me to upgrade, because I read the patches
that are in 4.1b2 and none of them is related to this problem.

Modem MultiTech MT1932ZDX and another class1 modem.

All is ok, but sometimes a fax sent to a number is sent to another one.
It's not a modem problem since it happens with 2 different modems.


The fax should go to number 0131867753, but it goes to 523 760315; this
is not systematic and not related to this particular number (the
following faxes for  0131867753 went to it)
The ATDT command reports a number, the remote CSI says it's another one.
Note that 523 760315 is the fax number of a job that failed just before
the job for  0131867753:
1) fax to 523 760315 (failed for BUSY)
2) fax to 0131867753 (that went to 523 760315)
3) retry of fax at point 1, 523 760315, that succesfully goes to 523
760315

The last time the sequence of events was very similar, the difference
was that the first call failed for "no carrier", instead of BUSY.

=== log of point 2:
Sep 23 08:24:37.27: [ 1937]: DELAY 300 ms
Sep 23 08:24:37.58: [ 1937]: <-- [16:ATDT0131867753@\r]
Sep 23 08:24:55.12: [ 1937]: --> [7:CONNECT]
Sep 23 08:24:55.12: [ 1937]: --> [2:OK]
Sep 23 08:24:55.12: [ 1937]: DELAY 300 ms
Sep 23 08:24:55.43: [ 1937]: <-- [9:AT+FRH=3\r]
Sep 23 08:24:55.44: [ 1937]: --> [7:CONNECT]
Sep 23 08:24:55.77: [ 1937]: --> [2:OK]
Sep 23 08:24:55.77: [ 1937]: REMOTE CSI "0039 523 760315"
Sep 23 08:24:55.77: [ 1937]: DELAY 300 ms
Sep 23 08:24:56.08: [ 1937]: <-- [9:AT+FRH=3\r]

=====

It's not that the other fax is misconfigured and report a wrong CSI, the
fax was really sent to another fax number.

I don't know c++, can someone tell me if the variable containing the
value displayed in ATDT is the same that HylaFAX uses to place the call?
It it's the same variable then HylaFAX is not the cause of this, but if
it uses 2 different variables then there may be a disalignment.

The 2 faxes were in a group of 10 faxes that were injected to sendfax
through a loop
do i = 1 to 10:
sendfax -n -d number[i] file[i]
end

Maybe sendfax may get confused if it receives too faxes at the same
time?

The customer in _angry_, if necessary I can send you the complete logs.

Thanks.

-- 
giulioo@tiscalinet.it