Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[lethosor]

It's pretty unlikely that those would contain anything, since keybinding only prints to the console.

[Max™]

Well, yeah, but worth a check in case there was something else going on, and useful generally to know where they are.

[lethosor]

The "string too long" exception might end up in stderr.log when you run keybinding, although I'm not sure if it would on Windows. It seems to be getFocusString, anyway, since that's the only non-trivial thing keybinding uses, but I can't imagine why that would crash with a specific DFHack screen.

[Ygvir]

@Max™ - Thanks for giving me the log files.

@lethosor - as you expected there isn't anything interesting in any of the files... but now we know that is the case.

[Lamp]

Hey guys,

I'm sure this gets asked a lot but how far out are we from a stable build for 0.43.05?

Thanks for all your hard work!

[Max™]

The alpha4 is pretty stable, just missing a few symbols and a couple aren't aligned properly.

[lethosor]

Which ones aren't aligned properly?

Also, there are a couple random crashes I haven't been able to track down yet, including the manipulator crash on 64-bit Windows and a box-select crash on 64-bit Linux.

[Max™]

Which ones aren't aligned properly?

Also, there are a couple random crashes I haven't been able to track down yet, including the manipulator crash on 64-bit Windows and a box-select crash on 64-bit Linux.

So, I hadn't thought much about this since originally I just attributed it to being on a personally compiled super-alpha develop branch of dfhack, but now with so many of the globals located I was trying to get some of my scripts back in order and ran into a problem with artifake not using the right value for artifact_next_id, and it seems to be because there isn't a value that makes sense for it at all?

Spoiler (click to show/hide)

I know that it should be like 17980 or so, not -159036100, and there are similar value issues for item_next_id and others seen there. The same problem happens with the alpha4 release so I'm rather confused that it otherwise works fine despite this.

Well, those addresses must be wrong. I'm not sure how exactly they were located, but as long as they're within the section of memory that contains globals, which they probably are, you'll get unpredictable values instead of crashes.

Incidentally, it could be easier to locate the correct address if you know exactly what those values should be, as long as they're not particularly low.

Code: [Select]

activity_next_id -1802810881 > 22785
artifact_next_id -159036100 > 44417
item_next_id 2055948296 > 1903806 
unit_next_id -2079390469 > 71293
hist_event_next_id 1149847552  > 720562
hist_figure_next_id -11891208 > 76043
squad_next_id -2092433408 > 110Those should be right, I just checked a few which changed while I was checking (hist_event_next_id should have been and the newest entry in events was indeed 720562) so yeah, they're pretty whacky, but that would be why launch/advfort crash me and artifake just spits errors but so many other scripts work fine.

Those ones for sure aren't aligned properly.

[lethosor]

Oh, the addresses aren't right. I thought you meant the structures were badly aligned (e.g. incorrect world field sizes). That's another thing we need to fix. I'll open Github issues for these if I can't find any.

[Max™]

I had intended to but forget what happened, I think I was on the tail end of a couple of days awake and ended up passing out.

[Quietust]

Using a few prepared saves, I've located all of the *_next_id globals for 64-bit Linux. There are still a few others to locate, though they hopefully shouldn't be too difficult.

[lethosor]

Link to new symbols.xml
Dropping that into your hack folder should do the trick.

[Quietust]

I believe I've fixed the Manipulator problem - it was a bug in our RTTI parser for 64-bit Windows code.

[Max™]

Using a few prepared saves, I've located all of the *_next_id globals for 64-bit Linux. There are still a few others to locate, though they hopefully shouldn't be too difficult.

A goddamn scholar and gentleman, truly.

[Roses]

Does anyone have a list of things in the raws that don't translate directly into the game? These are all the two I know of,

PHYS_ATT_RANGE is capped at 5000 when read into the game, but creatures can have a higher value
BODY_SIZE also gets capped (although I don't remember the exact value)

Anyone know if there are others?