PharoInbox/SLICE-M7343-Socket-Rollover-bug-nice.1
After http://bugs.squeak.org/view.php?id=7343 Socket timeout can suffer rollover problem (Happens on server with high load and uptime>12 days) I did not load Andreas change directly 1) He still uses questionnable (deadline - Time millisecondClockValue) which is potentially >0 forever 2) because of above choice, he protected with unecessary 500msecs busy loop (well necessary in this implementation, but not in absolute) 3) he does not answer true when ok: HTTPSocket requires this... I moved the *Until: family message to a deprecated category, but did not send (self deprecated: ...) I corrected in image senders. Do not remove now, they might be used by packages. _______________________________________________ Pharo-project mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project |
http://code.google.com/p/pharo/issues/detail?id=778
On Apr 28, 2009, at 10:25 PM, Nicolas Cellier wrote: > PharoInbox/SLICE-M7343-Socket-Rollover-bug-nice.1 > > After http://bugs.squeak.org/view.php?id=7343 > Socket timeout can suffer rollover problem > (Happens on server with high load and uptime>12 days) > > I did not load Andreas change directly > 1) He still uses questionnable (deadline - Time millisecondClockValue) > which is potentially >0 forever > 2) because of above choice, he protected with unecessary 500msecs > busy loop > (well necessary in this implementation, but not in absolute) > 3) he does not answer true when ok: HTTPSocket requires this... > > I moved the *Until: family message to a deprecated category, but did > not send (self deprecated: ...) > I corrected in image senders. > Do not remove now, they might be used by packages. _______________________________________________ Pharo-project mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project |
BEWARE: PharoInbox/SLICE-M7343-Socket-Rollover-bug-nice.1 is bogus
because there were subtle difference between wait*Until: family and wait*For: family. The former did silently return false on timeout/closed connection. The later did signal an error... USE PharoInbox/SLICE-M7343-Socket-Rollover-bug-nice.2 2009/4/29 Stéphane Ducasse <[hidden email]>: > http://code.google.com/p/pharo/issues/detail?id=778 > > > On Apr 28, 2009, at 10:25 PM, Nicolas Cellier wrote: > >> PharoInbox/SLICE-M7343-Socket-Rollover-bug-nice.1 >> >> After http://bugs.squeak.org/view.php?id=7343 >> Socket timeout can suffer rollover problem >> (Happens on server with high load and uptime>12 days) >> >> I did not load Andreas change directly >> 1) He still uses questionnable (deadline - Time millisecondClockValue) >> which is potentially >0 forever >> 2) because of above choice, he protected with unecessary 500msecs >> busy loop >> (well necessary in this implementation, but not in absolute) >> 3) he does not answer true when ok: HTTPSocket requires this... >> >> I moved the *Until: family message to a deprecated category, but did >> not send (self deprecated: ...) >> I corrected in image senders. >> Do not remove now, they might be used by packages. > > > _______________________________________________ > Pharo-project mailing list > [hidden email] > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > _______________________________________________ Pharo-project mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project |
Free forum by Nabble | Edit this page |