[asterisk-dev] [Code Review]: Extend the max number of callgroups/pickupgroups
Matt Jordan
reviewboard at asterisk.org
Thu Aug 2 10:23:24 CDT 2012
> On Aug. 11, 2011, 2:07 a.m., Olle E Johansson wrote:
> > I don't think we should have any compile-time-options or a fixed limit. If we fix this, we should have named pickup groups that can have other attributes as well in the future, possibly using ast_objects. Using bitmaps and numeric matching was something useful when CPUs had less power than today, when people have dual core in their pockets. This patch is just a short-term patch, not the solution to the problem for me.
>
> amorsen wrote:
> Personally I believe that the old callgroup/pickupgroup system should be replaced completely by Pickup(@PICKUPMARK) and only remain for backwards compatibility for a while.
>
> However, you can have multiple callgroups while one channel can only be assigned to one PICKUPMARK (fortunately Pickup handles 123 at PICKUPMARK&134&PICKUPMARK). If it was possible to add multiple PICKUPMARKs somehow, it would be possible to deprecate callgroups/pickupgroups with no loss of functionality. If we can find someone to do the coding, I can probably find a sponsor for the work :)
>
> In the meantime it is difficult to argue against code which is already written. The only objection I have is why not use long or long long? int is 32-bit on all the 64-bit systems supported by Asterisk right now, whereas long is 64-bit on those except on Windows. Long long is 64-bit everywhere even on Windows, at the cost of software emulation on 32-bit.
>
>
>
> Thomas Arimont wrote:
> The suggested fix on basis of named groups without a fixed limit is going to make it into Asterisk 11 now.
> Please see https://reviewboard.asterisk.org/r/2043 'Extended maximum number of pickupgroups (callgroup/pickupgroup)'
As Thomas said, the named groups in review 2043 provide a better implementation of pickup groups and meet the intent that Olle described. I'm going to go ahead and close this review out.
- Matt
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/466/#review4028
-----------------------------------------------------------
On June 22, 2010, 1:22 p.m., raarts wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/466/
> -----------------------------------------------------------
>
> (Updated June 22, 2010, 1:22 p.m.)
>
>
> Review request for Asterisk Developers.
>
>
> Summary
> -------
>
> Currently the number of callgroups/pickupgroups is limited to 63, because the variable type that holds the group bits is a long long.
>
> Here's a patch that extends this variable to an arbitrary-length (determined at compile time) int array. It currently sets the max number of groups to 1024.
> It implements this for all channels, but I only tested SIP and DAHDI.
>
> Some people might object to implementing this as an array of ints, because it introduces a loop. But I think call pickups are not done often enough for this to become a problem.
>
> See https://issues.asterisk.org/view.php?id=16640
>
>
> Diffs
> -----
>
> /trunk/apps/app_dumpchan.c 271908
> /trunk/channels/chan_agent.c 271908
> /trunk/channels/chan_dahdi.c 271908
> /trunk/channels/chan_gtalk.c 271908
> /trunk/channels/chan_jingle.c 271908
> /trunk/channels/chan_mgcp.c 271908
> /trunk/channels/chan_misdn.c 271908
> /trunk/channels/chan_sip.c 271908
> /trunk/channels/chan_skinny.c 271908
> /trunk/channels/chan_unistim.c 271908
> /trunk/channels/misdn_config.c 271908
> /trunk/doc/asterisk-mib.txt 271908
> /trunk/funcs/func_channel.c 271908
> /trunk/include/asterisk/channel.h 271908
> /trunk/main/channel.c 271908
> /trunk/main/cli.c 271908
> /trunk/main/features.c 271908
> /trunk/res/snmp/agent.c 271908
>
> Diff: https://reviewboard.asterisk.org/r/466/diff
>
>
> Testing
> -------
>
> This patch has been in our asterisk 1.2 for 3 years now without problems. I ported it over to 1.6.2 and tested it there. The patch uploaded here is against trunk which is currently being tested by a colleague of mine, but they are all basically the same patch.
>
>
> Thanks,
>
> raarts
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20120802/7c6daa53/attachment.htm>
More information about the asterisk-dev
mailing list