[asterisk-dev] 1.4: Res_smdi and chan_dahdi

[Olle E. Johansson]

I think I've discovered an issue with chan_dahdi depending on res_smdi in 1.4. I can't load res_smdi at the cli, and thus can not load chan_dahdi.

/O

From the IRC this afternoon:

oej:
Afternoon
[14:26]oej:
Dear sirs, I have a problem with Asterisk.
[14:26]oej:
chan_dahdi depends on res_smdi
[14:27]oej:
But if I load res_smdi in the cli, it refuses to load
[14:27]oej:
and chan_dahdi can't be loaded
[14:27]oej:
res_smdi reads the empty config file (the default sample) and unloads because there's nothing to do
[14:27]oej:
which makes it impossible to load chan_dahdi
[14:28]kpfleming:
that seems broken
[14:29]kpfleming:
the smdi dependency should be optional, not mandatory
[14:44]jmls joined the chat room.
[14:51]TheDavidFactor joined the chat room.
[15:01]jeffg joined the chat room.
[15:03]oej:
yes, it seems very broken to me.
[15:05]oej:
We have a "depends on" for res_smdi in chan_dahdi, not "can use"
[15:06]oej:
There are no "#ifdef" in chan_dahdi for smdi
[15:06]oej:
so this is very, very, very broken
[15:09]kpfleming:
well, it doesn't require #ifdef for optional usage, because the API could live in the asterisk core (like some already do) and just be provided by res_smdi
[15:10]kpfleming:
in fact, i think that's how it is supposed to be now... which branch is this?
[15:11]JT left the chat room. (Ping timeout: 252 seconds)
[15:26]kpfleming left the chat room. (Ping timeout: 246 seconds)
[15:29]jsmith-away is now known as jsmith.
[15:35]oej:
Yes, we could have stubs, right
[15:35]oej:
I'm using good ol' classic 1.4
[15:35]oej:
I haven't tried loading at start time
[15:35]oej:
Discovered this by loading after startup at the CLI