> On Jan. 8, 2015, 9:13 p.m., Mark Michelson wrote: > > branches/13/res/res_pjsip.c, lines 43-45 > > <https://reviewboard.asterisk.org/r/4311/diff/1-2/?file=70130#file70130line43> > > > > What caused these dependencies to be added? > > Kevin Harwell wrote: > pjsip_options uses sorcery_memory with regards to qualify/contact_status > and contacts themselves are stored in the astdb (location.c) > > Joshua Colp wrote: > I think res_sorcery_memory is fine being a dependency because it is > actually required by stuff internally. > > res_sorcery_astdb on the other hand is only required if you are using the > default of storing within astdb - it's runtime configurable to be different. > > George Joseph wrote: > astdb is the persistent store for contacts. res_pjsip will fail to load > if res_sorcery_astdb isn't loaded. > > > Joshua Colp wrote: > In a default configuration. > > George Joseph wrote: > True but my point in asking for MODULEINFO to be updated was that it's > not obvious that res_pjsip needs astdb to load in most cases. In fact, if > you're setting up realtime what would even trigger you to override contact > since contact isn't an object you normally define in pjsip.conf? So unless > you actually looked in location.c, contact is going to need astdb.
I'm fine with making it a dependency since subscription persistence does pretty much require it. Ultimately though since this is only build time dependencies and not runtime it's possible to still run into a problem. - Joshua ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviewboard.asterisk.org/r/4311/#review14126 ----------------------------------------------------------- On Jan. 9, 2015, 3:37 p.m., Kevin Harwell wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviewboard.asterisk.org/r/4311/ > ----------------------------------------------------------- > > (Updated Jan. 9, 2015, 3:37 p.m.) > > > Review request for Asterisk Developers. > > > Bugs: ASTERISK-24485 > https://issues.asterisk.org/jira/browse/ASTERISK-24485 > > > Repository: Asterisk > > > Description > ------- > > The res_pjsip module was previously unloadable. With this patch it can now be > unloaded. > > This patch is based off the original patch on the issue by Corey Farrell with > a few modifications. Removed a few changes not required to make the module > unloadable and also fixed a bug that would cause asterisk to crash on > unloading. > > > Diffs > ----- > > branches/13/res/res_pjsip/pjsip_outbound_auth.c 430294 > branches/13/res/res_pjsip/pjsip_options.c 430294 > branches/13/res/res_pjsip/pjsip_global_headers.c 430294 > branches/13/res/res_pjsip/pjsip_distributor.c 430294 > branches/13/res/res_pjsip/pjsip_configuration.c 430294 > branches/13/res/res_pjsip/location.c 430294 > branches/13/res/res_pjsip/include/res_pjsip_private.h 430294 > branches/13/res/res_pjsip/config_transport.c 430294 > branches/13/res/res_pjsip/config_auth.c 430294 > branches/13/res/res_pjsip.c 430294 > branches/13/main/stasis_message_router.c 430294 > > Diff: https://reviewboard.asterisk.org/r/4311/diff/ > > > Testing > ------- > > Made it so res_pjsip was the only pjsip module loaded and then issued an > unload and noted it unloaded successfully (also loaded/unloaded it several > times from the CLI). Also when loaded and with REF_DEBUG enabled issued a > "core stop gracefully" and made sure there were no ref leaks for the module. > > Also tested unloading with other dependent pjsip modules loaded and noted > that the module would not unload (as it should since dependencies are > currently loaded). And then shutdown asterisk and made sure it did not crash > or anything. > > Started asterisk with nominal and off nominal module and pjsip configurations > to make sure things behaved appropriately (no crashes and such) and then > attempted to, or successfully unload the res_pjsip module. Also made sure > Asterisk continued to shutdown without incident. > > > Thanks, > > Kevin Harwell > >
-- _____________________________________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- asterisk-dev mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev
