[hylafax-devel] Re: Modem Reset Commands

[Dmitry Bely <dbely@mail.ru>]

Nico Kadel-Garcia <raoul@adsl-151-203-22-73.bellatlantic.net> writes:

> > > > Seems we have finally reached consensus :-) So you are agree now with
> > > > explicit ModemSoftResetCmd (ATZ), issued *before* any other initialization
> > > > activity? Who else is against it? :-)))
> > > 
> > > Works for me. Can we gracefully make it something like "ATZ\nAT&F" in order
> > > to restore the modem to the factory default, or do we have a separate
> > > AT&F slipped in elsewhere?
> > 
> > Personally I am against it. I like to store some site-specific values
> > (e.g. TX level, required for the specific telco line) and don't write them
> > to init strings of each and every application. Also take into account that
> > Hylafax sets *all* parameters needed for correct faxing itself. Anyway, if 
> > someone really needs AT&F, he/she can easily add it to ModemResetCmds
> > string.
> 
> Sigh. That's the problem: you are storing non-modem specific information
> and relying on it for every session, then you turn around and say
> HylaFAX sets all the stuff needed for faxing.

There are

1. Fax specific settings (AT+F commands)
2. Common fax/data settings (command echo, auto-answer etc...)
3. Model specific settings (flow control, bit order ...)
4. Line/site specific settings (pulse interdigit time, signal transmit
level, min/man transmit/receive speed, ...). Can be stored in the profile
(ATZ) and/or done via explicit ModemResetCmds.

(executed in the reverse order: 4-3-2-1)

Hylafax itself sets (1),(2) and (3) (if correct modem config is
supplied). All this does not depend on modem's profile (ATZ) at all. Is
this enough for faxing? Yes. Will the modem work better with site specific
sittings (4)? Yes also. Is this still not clear?

> HylaFAX also does data, and other modems can and do use the phone. If I
> swap the modems from one system to another, it is *possible* for someone
> else using it to have it in a really unfortunate state that will be
> restored by ATZ and the power cycling.
> 
> > > My concern is the potential misconfiguration
> > > by other applications, restored by an ATZ.
> > 
> > ATZ state is automatically restored
> > a) after power-on
> > b) after DTR drop
> > 
> > So if your (or someone's else) installation works now, it will also work
> > after explicit ATZ in the beginning. Nothing will change. What should we
> > test? (I also reply here to your parallel message)
> 
> You're clearly missing the point. Here's a potential timeline.
> 
> 	I install and run HylaFAX on my brand new dual-boot
> 	Linux machine with Zoom modem. It runs, I'm happy, I work from a
> 	known good state
> 
> 	I install SCO Xenix on another partition (yecchh!!!). Zoom
> 	sends me the ghods-awful init strings of death they used to
> 	send out, completely uselessly, and I try to get UUCP working.
> 	I store the configuration, which is just plain wrong for
> 	anything useful, and it never works correctly.

What if you simply remove the disk partition with Hylafax or delete some
config files? The problem of the same kind.

> 	I reboot to Linux, thankful to escape with my life, and
> 	ATZ is now restoring my modem to some state invented by
> 	the drunkard with a Ouija boards called "UNIX tech support"
> 	at Zoom. I am now dead in the water until I do an AT&F, possibly
> 	followed by an AT&W, to get ATZ working the way I expect again.

Does the current release of Hylafax do AT&F? Answer: *NO*. Did you
complained before that the initial state of the modem was not known when
Hylafax started (for the *current* release)? Answer: *NO*. Did *anybody*
complained? Answer: *NO*.

> I believe that the behavior of ATZ is to some extent modifiable by the
> registers. Working "on any reasonable installation" is not enough: most
> sites don't folks as smart as us to go in and do this kind of resetting
> by hand.

I do not do *any* additional assumptions comparing to the current
Hylafax. Seeing extra ATZ just imagine that you have switched your modem
off and on. Is this also not acceptable? :-)

Hope to hear from you soon,
Dmitry



____________________ HylaFAX(tm) Developers Mailing List ____________________
 To unsub: mail -s unsubscribe hylafax-devel-request@hylafax.org < /dev/null