Notes |
|
(0007229)
|
AVS
|
2017-08-14 14:55
|
|
What are actual type names of i->second.type() and typeid(VectorizedObjectInfo<T, U>)? |
|
|
(0007260)
|
SXX
|
2017-09-04 19:30
|
|
I uploaded the log and also get report that problem also present on FreeBSD with Clang 4.0. |
|
|
(0007261)
|
AVS
|
2017-09-04 20:38
|
|
I see from log that VectorizedObjectInfo is probably valid Creature serilizer, but what is `i->second.type()`? |
|
|
(0007262)
|
AVS
|
2017-09-04 20:41
|
|
I suspect typeinfo-related bug. What if we replace typeid comparing with type name comparing - or even remove that assertion. |
|
|
(0007265)
|
SXX
|
2017-09-06 16:42
|
|
i->second-> content is:
content boost::any::holder<VectorizedObjectInfo<CCreature, CreatureID> > * 0x113f7e370
0x0000000113f7e370
So I suppose it's it. |
|
|
(0007266)
|
AVS
|
2017-09-06 16:45
|
|
Then just remove assertion, this is yet another typeinfo issue on osx |
|
|
(0007267)
|
SXX
|
2017-09-06 16:47
|
|
Question is: why it's only occur in newer Xcode.
Also I guess this could easily be in new Clang on other platforms as well. |
|
|
(0007268)
|
AVS
|
2017-09-07 10:36
|
|
We should replace all typeinfo comparisons with name comparison like in TypeComparer |
|