OpenMW 0.28.0

[dEnigma]

I have an old English version here. But I don't have the Expansion, so the highest version of Morrowind I can test is 1.2.0722 (=latest patch for morrowind without addons), which loads my non-NULL-terminated morrowind.esm just fine.

edit: when I add a null to the end I get a warning message but it still seems to work fine apart from that

[Zini]

Even if morrowind engine can't support both styles, what holds us from doing it?

Quoting myself:

reader: Strip terminating 0 from strings.

Which means we are supporting loading content files of both styles.

But we still have to decide what style OpenCS uses when writing and the reasonable thing here is the no-terminator flavour.

[Greendogo]

Why did Bethesda make the change to begin with? Is there a benefit?

[Zini]

Nope. Entirely redundant.

[sirherrbatka]

Forgot about morrowind compatibility

Anyway ─ I think that getting rid of the null-terminated strings is a logical step. However, Bethesda surely had any hidden intention when going for null-terminated strings. Maybe this is some sort of workaround? Also: will this break compatibility with third party tools like enchanted editor? I know that they are going to be obsolete when OpenCS start up but this is not a reason to break it.

[Zini]

I think it won't cause any problems, since 3rd party tools must already deal with both flavours.

[Zini]

Did look around a bit more in my Morrowind.esm file. It seems most of the strings actually are 0-terminated. There are only some cases, where the 0 is missing. The gives us an additional complication. What if MW (and 3rd party tools) require the terminating 0 here. We haven't actually seen any copy of Morrowind.esm, where no strings are 0-terminated. Right?

Maybe we should leave the CString function in, but only use non-0-terminated string variant for new records. That way we are on the safe side, even if it is less consistent.

[psi29a]

https://code.google.com/p/bullet/issues/detail?id=371

^--- hahahah

Need to look into this soon, but how did this issue get >200 stars in no time?

The normalsInWorld is set to true for btScaledBvhTriangleMesh.

OpenMW: it seems that you have some other issue related to convexCast normals, can you create a reproduction case in the Bullet/Demos/RaytestDemo and file a new issue?

^-- I think it is your turn Chris. At least they are listening to us! Cheers!

[sirherrbatka]

http://dl-common.cvs.sourceforge.net/vi ... orrow/esm/

The string class is the EsmSubNameFix.h

From Dave Humphrey

[Mistahtokyo]

Not sure how many people have noticed yet, but it seems the bullet issue has been resolved!