Print

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

[Saiko Kila]

Anyone else has this problem with Win x64 build (0.44.12-alpha1):
 When using "die" to exit DF, the DF's window disappear, but the DFHack's window still hangs there, and the whole Dwarf Fortress.exe process doesn't exit too? I need to kill it manually with process explorer.

[clinodev]

Anyone else has this problem with Win x64 build (0.44.12-alpha1):
 When using "die" to exit DF, the DF's window disappear, but the DFHack's window still hangs there, and the whole Dwarf Fortress.exe process doesn't exit too? I need to kill it manually with process explorer.

With Win x64 build (0.44.12-alpha1), and the 0.44.11 alpha as well, the DFHack window does hang for me for longer than expected, closing on its own after 15s or so. I haven't had to end the process (after noting it indeed did eventually close the first few times, now I just go on with my business. I only use "die" when testing quick things in disposable worlds.) Windows 10.

[Saiko Kila]

Anyone else has this problem with Win x64 build (0.44.12-alpha1):
 When using "die" to exit DF, the DF's window disappear, but the DFHack's window still hangs there, and the whole Dwarf Fortress.exe process doesn't exit too? I need to kill it manually with process explorer.

With Win x64 build (0.44.12-alpha1), and the 0.44.11 alpha as well, the DFHack window does hang for me for longer than expected, closing on its own after 15s or so. I haven't had to end the process (after noting it indeed did eventually close the first few times, now I just go on with my business. I only use "die" when testing quick things in disposable worlds.) Windows 10.

Thank you, this prompted me to check it out on Win 10. I have a dual boot PC with Windows Vista and Windows 10.

On Windows 10 it works as you said: first time in hangs for several seconds, subsequently it stops for only a second or two (so overall: correctly). I thought that problem may be due to a difference in antivirus (Avast on Vista, Kaspersky on Win10), but disabling Avast on Vista apparently didn't help. Something has changed between 0.44.11-alpha1 and 0.44.12-alpha1, which prevents the "die" to work in full on my system.

[plenum]

Hey!
Have anyone tried to copy burrow? Can't find anything on that... I wonder how hard is it do in script?

[Meph]

Just wanted to say that you guys are amazing and keep up the good work. :-) Today is just waiting for your updates before he releases the next version. 

[iceball3]

Hey, all this new update hype made me wanna to drop by and say, thank you all who work on DFHack between updates. <3
No rush, just saying that I really appreciate the work y'all have done to get this stuff going well!

[Rumrusher]

it seems the new dfhack adds a new folder called 'included' which I'm wondering if that suppose to add something?

[Saiko Kila]

it seems the new dfhack adds a new folder called 'included' which I'm wondering if that suppose to add something?

This "include" directory looks like a part of source code that was included mistakenly.

This may be related to my problem, which I described a few posts back - that the console window doesn't go away after the "die" command, and this holds the whole Dwarf Fortress.exe process (which must be killed by external means).

I have checked all released versions of DFHack for 0.44.11, and the last one which didn't have the problem was DFHack 0.44.11-alpha1.
The next version, DFHack 0.44.11-beta1 was bugged (console not initialised at all), and all the subsequent versions, starting with DFHack 0.44.11-beta2, hang on "die" on my system. They also include these directories "include" and "lib", coincidentally or not.

Something was changed with the bugged version, and this change does affect my system, even though it doesn't affect Windows 10. It looks like the console doesn't quit cleanly.

[lethosor]

The "include" and "lib" folders are due to jsoncpp changes and were included accidentally. They won't be in future releases. They're unused and are entirely unrelated to your console issues.

I'd recommend reading the release notes. The ones for beta2 list a few significant console changes, which are the reason for any new console bugs, like the one in beta2. We're not hiding anything - in fact, the announcement on GitHub explicitly asked people to look for and report console issues.

Anyway, "die" calls _exit() and nothing else, which is defined to kill the current process immediately. As far as I know, it's not even possible for it to hang, so I have no idea what's going on there. Can you report this on GitHub so it won't be forgotten? https://github.com/DFHack/dfhack/issues/1356

[Quietust]

Anyone else has this problem with Win x64 build (0.44.12-alpha1):
 When using "die" to exit DF, the DF's window disappear, but the DFHack's window still hangs there, and the whole Dwarf Fortress.exe process doesn't exit too? I need to kill it manually with process explorer.

I'm actually seeing a totally different problem - with the 32-bit version on Windows 7 x64, it outright crashes if I type "die" in the console, but if I exit the game cleanly, it works fine. The 64-bit version doesn't seem to be affected.

[Rose]

I've had hangs from pressing die in the console, but reproducing them has been tricky.

[Saiko Kila]

I've checked threads of bugged process. The function which hangs in my case is called _crt_at_quick_exit (it's a kernel function), which has state "Wait, executive". This means it waits for thread scheduler, which never shows interest in the thread. I can speculate that memory management or garbage collection mechanisms of Windows 10 take care of this, while ones from Vista can't, and from Win7 maybe partially can.

I'm not a specialist in this matter, but the quick exit function has some precautions, like registering I/O and generally functions to be flushed when it's called, so it's not exactly fire and forget. Maybe they are not registered properly.

[lethosor]

Suokko and BenLubar traced it down to _exit and _Exit not doing their job properly on Windows, basically (the standard specifies that _Exit shouldn't call global destructors, but it is). Suokko produced a fix that might work: https://github.com/DFHack/dfhack/pull/1357

[lethosor]

Here is a test build that should fix the Windows "die" issues. Can some of you test it?

[iceball3]

Probably not the platform you're looking for tests, but I managed to use "die" successfully ingame using that new 64 bit build.