yang686526

odgsbasemodel:nmodified(odgidrawable* pmodified, oddbstub*
odgsbasemodel:nmodified(odgidrawable* pmodified, oddbstub* parentid)
what does this function do? how is it implemented? can i get the source for this function?
it doesn't actually crash here... however if i avoid invalidating the gs cache then i don't get my crash in device::update where my call stack has the spatial index dll at the top.
nearer!
wvk,
thanks for the code.
commenting this line:
pcache->invalidate(pparentcache, 0,0);
alleviates my crash. so...now what is the odgscache::invalidate doing? i'm assuming its using polymorphism to call the odgsentitynode::invalidate. somehow... this code does something different (than an add) which causes the device::update to crash.
tyler
i've noticed that the extents call on the odgientitynode returns a different value than that of the getgeomextents call on the odgidrawable. i'm wondering if that is somehow fouling up the spatialindexing (each odgientitynode is spatialindexed). any idea how they would become out of synch... or how to synchronize them?
wvk,
after doing some further research... im pretty sure the crash occurs deep in the odgsbasevectorizedevice::update call. im sure that it has to do with spatialindexing of the odgsentitynode that i've called invalidate on.
questions:
1. does the odgsentitynode::invalidate just set a flag? or does it do some calculation?
2. what is the difference between the odgsnode extents and the odgidrawable extents? is the odgsnode extents calculated in the update?
3. can i get the odgsbasevectorizedevice::update code?
vladimir,
i've done more investigation into this issue.
it consistently reproduces itself in dd_spatialindex.2.04_8. dll. it appears to get caught in either an infinite loop or a recursive function that allocates 12 bytes per iteration. im able to set a break point in _heap_alloc_dbg, and by watching the call stack i see two function addresses toggle back and forth.
i've tried everything i can think of as a work around using the odgsnode and odgidrawable ptrs from the onmodified function...but can't seem to hack around this issue.
i'd really appreciate the odgsbasevectorizedevice::update source and the spatial indexing dll source so that i can trace the issue in code. let me know if you can provide them to me... or who i should be asking this for.
thanks,
tyler
tillerstarredwards