Originally created by: bre... (code.google.com)@gmail.com
Originally owned by: timoor... (code.google.com)@gmail.com
What steps will reproduce the problem?
Random segfaults may occur since properties are modified/updated/merged in multiple threads (update traversal, main loop, ...) though the property tree implementation isn't (fully) thread-safe.
Not easily reproduced and doesn't happen too often. Maybe more likely to happen while loading many models (playing MP / moving between scenery areas...).
Attachement shows a crash stack-trace. Shows two threads working concurrently on properties: one is merging property trees while the other crashes on SGPropertyNode access.
View and moderate all "codetickets Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Tickets"
Originally posted by: bre... (code.google.com)@gmail.com
(No comment was entered for this change.)
Cc: zakalawe
View and moderate all "codetickets Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Tickets"
Originally posted by: bre... (code.google.com)@gmail.com
Obviously not a major issue for now (I rarely see any issues).
Labels: Priority-Low
View and moderate all "codetickets Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Tickets"
Originally posted by: bre... (code.google.com)@gmail.com
(No comment was entered for this change.)
Status: Stalled
View and moderate all "codetickets Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Tickets"
Originally posted by: wonco...@googlemail.com
Not sure if this is related to the property tree, but I have lately been seeing 2-3 segfaults per 10-20 FG starts that were obviously related to multithreading issues, i.e. re-starting FG afterwards, would "solve" the problem - without having touched ANY settings at all.
Most of these segfaults happened while starting FG, i.e. during initialization.
View and moderate all "codetickets Discussion" comments posted by this user
Mark all as spam, and block user from posting to "Tickets"
Originally posted by: zakalawe@mac.com
Closing this out because assorted threading-related property crashes are now fixed (in 3.0), e.g. we synchronise & cache material validity computations and effect/technique realisation.
Status: WontFix