[asterisk-dev] [Code Review] dial by name in chan_dahdi

[rmudgett at digium.com]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/535/#review1621
-----------------------------------------------------------


This is definitely going to be a useful feature!

FYI: You might want to look at the asterisk/team/group/CCSS branch.  The changes to dahdi_request() here conflict with the dahdi_request() parse code refactoring done in the CCSS branch.


/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3627>

    Please add a doxygen function header comment.  You can use the dahdi_new_pri_nobch_channel() comment header as a template.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3633>

    This ast_log() message is unconditional.  It really should be a debug or verbose type message if you want to keep it.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3626>

    Please add a doxygen description of subdir to the function header comment.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3635>

    Please add a doxygen function header comment.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3630>

    The ast_log() here is essentially redundant with the message generated by the calling location and can be removed.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3629>

    Shouldn't the built path be from root like it is in device2chan()?
    "/dev/dahdi/%s/%d"



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3631>

    This ast_log() message is unconditional.  It really should be a debug or verbose type message if you want to keep it.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3632>

    Update this comment to show the new dialing format allowed.



/trunk/channels/chan_dahdi.c
<https://reviewboard.asterisk.org/r/535/#comment3634>

    Please add a doxygen function header comment.


- rmudgett


On 2010-03-01 16:24:55, Tzafrir Cohen wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/535/
> -----------------------------------------------------------
> 
> (Updated 2010-03-01 16:24:55)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> Currently a name of a channel in chan_dadhi must (with some exception of trunk-groups syntax) rely on DAHDI channel numbers. Sadly channel numbers depend on the specific order of registration. Changing that seems to be quite difficult.
> 
> Thus we try to use the original ope-by-device-file (rather than indert open through /dev/dahdi/channel), and allow files to reside in subdirectories of /dev/dahdi .
> 
> Specifically, the file name must still be a number (to allow using the same ranges syntax). However it should be preceded with the string '<dirname>!' .
> 
> For instance we have: /dev/dahdi/work/1 and /dev/dahdi/work/2 that are the devices <196,9>, <196,10> (DAHDI channels 9 and 10), I can refer to them in chan_dahdi.conf as:
> 
>   channel => work!1
>   channel => work!2
> 
> And likewise can dial through them using:
> 
>   Dial(DAHDI/work!1/123456)
> 
> 
> See also Oron's older asterisk-dev message at http://lists.digium.com/pipermail/asterisk-dev/2008-July/033805.html
> 
> 
> Current status: 
> 
> * Tested to basically work. 
> 
> TODO:
> * At the DAHDI side two major things are missing:
>   1. Device files for channel numbers > 249
>   2. More automated generation of device files.
> * Adapt better to Asterisk coding style.
> 
> 
> Diffs
> -----
> 
>   /trunk/channels/chan_dahdi.c 249755 
> 
> Diff: https://reviewboard.asterisk.org/r/535/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Tzafrir
> 
>