[Asterisk-bsd] Update on zaptel-bsd

[Oleksandr Tymoshenko]

John Todd wrote:
> [Digium hat on, but I don't feel like re-subscribing to this group 
> from my Digium account.  Replies to jtodd at digium.com please.]
>
> So, it would be great if the zaptel-bsd community could provide:
>
>   1) Two volunteers for keeping the *BSD portions of the code up to date.
>   2) FAQ, README, or other documentation on the distribution archive
>   3) Support for FreeBSD, OpenBSD, and Darwin in the package
>
> If that is the case, I'll further try to get this adopted as either 
> an add-on or basic portion of the zaptel/DAHDI distribution.  Again, 
> I can't say it will happen, only that I will try.
>   
    Thank you for reply, John. Unfortunately you're right and in current 
state zaptel-bsd has no
chance to make it to mainline. Moreover there is no point in it. Because 
it would duplicate all
functionality of linux version. In ideal world zaptel/DAHDI would have 
thin OS-dependent
"compatiblity layer" and major part of logic being OS-indepent. So it's 
this "compatibility
layer" that should be in mainline.  Even if it would contain only Linux 
parts it would make
porting and maintaining of zaptel/DAHDI infrastructure much easier for 
other OSes.
Porting process is relatively easy, what is *really* hard is 
maintenance. It's either one
ports every release from the scratch or tries to merge every svn commit 
to your modified
sources. Both ways are PITA unless you don't have a bunch of scripts 
making your life tolerable.

    Although in modern *NIX derivatives a lot of KPI differs virtually 
by types/function names,
arguments number and/or their order one can't separate this OS-dependent 
layer from the rest
of code without  having both versions before one's eyes or without 
digging into other OS'es KPI.
That's where we come to item #1 and item #2 of your summary. Current 
distribution even with
README would do no good. Much better would be a module ported to FreeBSD 
with a list
of thing that have been changed. It would allow Digium to make some 
estimates of efforts for
creating OS-independent layer and therefor to decide if they're 
interested in it. I can do it for
non-hw "zaptel" module and for one of the hw drivers. And yes, I 
volunteer to maintain *BSD part
of code if things would be done right (tm) :)

-- 
gonzo