[asterisk-bugs] [JIRA] (ASTERISK-26422) [patch] Force calendars to do new fetch after module reload

[Martin Tomec (JIRA)]

    [ https://issues.asterisk.org/jira/browse/ASTERISK-26422?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=234379#comment-234379 ] 

Martin Tomec commented on ASTERISK-26422:
-----------------------------------------

Hi,
I have tested this patch with ical calendars, and after each "module reload res_calendar.so" I have more calendars loaded. Old calendars are probably not unloaded:
{noformat}
 asttest*CLI> calendar show calendars
 Calendar             Type       Status
 --------             ----       ------
 test2                ical       busy
 asttest*CLI> module reload res_calendar.so
 Module 'res_calendar.so' reloaded successfully.
 asttest*CLI> calendar show calendars
 Calendar             Type       Status
 --------             ----       ------
 test2                ical       busy
 test2                ical       busy
{noformat}

> [patch] Force calendars to do new fetch after module reload
> -----------------------------------------------------------
>
>                 Key: ASTERISK-26422
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-26422
>             Project: Asterisk
>          Issue Type: Improvement
>      Security Level: None
>          Components: Resources/res_calendar/NewFeature
>    Affects Versions: 13.9.1
>            Reporter: Ludovic Gasc (Eyepea)
>            Assignee: Ludovic Gasc (Eyepea)
>            Severity: Minor
>
> Hi,
> We use a lot res_calendar, we are very happy with that, especially because you use libical, the almost alone opensource library that supports really ical format with all types of recurrency.
> Nevertheless, some features are missed for our business use cases.
> This first patch adds a new option in calendar.conf: fetch_again_at_reload
> Be my guest for a better name.
> If it's true, when you'll launch "module reload res_calendar.so", Asterisk will download again the calendar.
> The business use case is that we have a WebUI with a scheduler planner, we know when the calendars are modified.
> For now, we need to define 1 minute of timeout to have a chance that our user doesn't wait too long between the modification and the real test.
> But it generates a lot of useless HTTP traffic.
> I'm putting the very small patch in Gerrit.
> Have a nice Astricon.



--
This message was sent by Atlassian JIRA
(v6.2#6252)