OpenMW 0.29.0

[cc9cii]

Release version has the same issue.

But I noticed that a savefile that had only a few entries in it (because I aborted it during debugging) does not have the issue. I'm wondering now whether the file get over-written with junk later.

I'll step through the save part once more.

EDIT: debug save results in "vector subscript out of range" in ESM::SavedGame::save()

Code: Select all

 esm.write(&mScreenshot[0], mScreenshot.size());

[Zini]

Empty screenshot vector?

[cc9cii]

I backed out gus's suggsted change on page 9 of this thread and that particular issue doesn't happen any more.

I've discovered that the savefile generated inside the prison ship (debug or release version with int changed to uint32_t) does not have "0D0A" for the size of "PLCE". However the release version does have the issue when I save outside the Census office (I can't try the same with debug version due to slaughterfish).

I have no idea what is going on.

[gus]

did you try binary mode? Because it does work on my computer.

[WeirdSexy]

I don't see a pattern. There are some topics ("TOPI") with "0D0A0000" as the size while other topics are fine.

The one with odd size has string "ackground" which I've never seen in years of playing morrowind. So it doesn't seem to be consistent.

Um, just a thought, but might that be "background" with the 'b' cut off?

[cc9cii]

did you try binary mode? Because it does work on my computer.

sorry, please remind me how to try binary mode

[Zini]

std::ofstream stream (slot->mPath.string().c_str(),std::ios::binary);

[cc9cii]

So here is a crazy thing. I go through the chargen process right up to the final door. I then hit F2 and set chargenstate to -1. Save. Go outside the door and save again.

The save from inside is ok but the save from outside is not. (still none of them will load properly)

I'll try the binary mode now.



That worked! Thanks everyone!

[gus]

So if I sum up, there were 3 differents bugs:

1)screenshot
2)binary mod
3)this uint32 thing

But we need a proper fix for the screenshot. No wonder it was hard to debug.

[Zini]

There is also the size_t/int issue. And what is the state on the screenshot?