Changes between Initial Version and Version 1 of Ticket #2427, comment 4


Ignore:
Timestamp:
Sep 7, 2014, 7:23:06 AM (10 years ago)
Author:
historic_bruno

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #2427, comment 4

    initial v1  
    11The reported crash is certainly a bug, `CreateEventStream` shouldn't crash just because an empty vector is passed in. Though I'm not sure if this will ever happen in normal conditions.
    22
    3 Now that I'm able to test on Mavericks again, it seems the issues with Atlas are more complex. After fixing the one crash, there is another in Atlas' message passing system. In fact, `InitVfs` is supposed to be run with other init code in response to an `Init` message, but that never happens. It crashes at [source:ps/trunk/source/tools/atlas/GameInterface/GameLoop.cpp#L165 GameLoop.cpp:165]:
     3Now that I'm able to test on Mavericks again, it seems the issues with Atlas are more complex. After fixing the one crash, there is another in Atlas' message passing system (this one occurs when starting from the main menu too). In fact, `InitVfs` is supposed to be run with other init code in response to an `Init` message, but that never happens. It crashes at [source:ps/trunk/source/tools/atlas/GameInterface/GameLoop.cpp#L165 GameLoop.cpp:165]:
    44{{{
    55* thread #6: tid = 0x37d92, 0x0000000100908a30 pyrogenesis_dbg`MessagePasserImpl::Retrieve() [inlined] std::__1::deque<AtlasMessage::IMessage*, std::__1::allocator<AtlasMessage::IMessage*> >::front(this=0x00007fff5fbfd348) + 23 at deque:1733, stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)