openmw.org

Considering they're built with VS2017, using the new UCRT, it's a very tentative maybe. There's nothing explicitly preventing Windows XP support, but I can most definitely not guarantee that they'll work.

Do we have someone with an XP VM who's willing to maybe give them a try, to be able to answer one way or the other?

You have to install XP support as an extension in 2015 and explicitly enable it, so I doubt that the builds support XP. This topic started with a theoretical discussion between me and Psi which wasn't really about whether the builds we produce work on XP, but rather how old a Windows version a really determined person could compile OpenMW on and whether that would change with a C++ version upgrade.

And I see now that my build toolset has changed from the _xp one, so I guess XP support on these builds is basically guaranteed to not exist, used to build with the XP toolset though just because it didn't hurt to do so.

Zini wrote: ↑31 Oct 2018, 13:21 Anyway, here is the backtrace:

Judging by backtrace, your bodypart here is not initialized properly, so when you try to lowercase this ID, you can get an out of bounds error during string mutation.
Since no one managed to reproduce this issue, probably you will have to track it down in the debugger (or at least print a bodypart Id, so we can track down which bodypart causes this issue).

https://github.com/OpenMW/openmw/pull/2029

I hope it will help.

If that works, should it be cherry-picked into 0.45 so we can get our 200 issues back?

Can Windows build please update its version of OpenAL to squash this bug?

psi29a wrote: ↑08 Nov 2018, 10:06 If that works, should it be cherry-picked into 0.45 so we can get our 200 issues back?

Definely yes.

Waiting on Zini

Feature is not important enough to restart the RC-process once more. I say just go ahead and add the fix to 0.46.