On Tue, 2009-04-14 at 16:40 -0400, David Saint wrote: > Hi, > > I've noticed that in many places in sipX we store the result of > getSecsSinceEpoch() in an "int" rather than in a "long" even though > > sipXportLib/include/os/linux/OsDateTimeLinux.h defines the return > value as "unsigned long". > 32bit machine: int=32bits long=32bits time_t=32bits > 64bit machine: int=32bits long=64bits time_t=64bits > > This won't be a problem until the year 2038 when the 32bit register > overflows, but it would be nice to fix it before then! (at least for > the 64bit machine). > > Some of the units that would need to be changed to store time in > "long" rather than "int" are: > sipXcallLib/src/cp/SipPresenceMonitor.cpp > sipXcommserverLib/src/sipdb/SubscriptionDB.cpp > sipXcommserverLib/src/persist/SipPersistentSubscriptionMgr.cpp > > sipXproxy/lib/authplugins/NatTraversalAgent/NatTraversalAgentDataTypes.cpp > sipXpublisher/src/statusserver/SubscribeServerThread.cpp > sipXproxy/lib/authplugins/NatTraversalAgent/NatMaintainer.cpp > sipXpublisher/src/statusserver/SubscribeServerThread.cpp > sipXpublisher/src/statusserver/Notifier.cpp > sipXregistry/src/SipRegistrarServer.cpp > sipXtackLib/src/net/SipSubscriptionMgr.cpp > > Dale has mentioned that we would need to check for (1) negative value > storage (2) IMDB value/types (3) registration "update number" > calculation. > > Maybe I should raise an issue against sipx 4.2 release to make this > change?
Nice catch. File the issue. _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
