VOGONS

Common searches


0.62 bugs overview

Topic actions

  • This topic is locked. You cannot reply or edit posts.

Reply 21 of 46, by Guest

User metadata

Hello !

I've been using dosbox 0.61 with a superbox 3 (playstation pad adapter recognized as a Microsoft Joystick with 4 axes and 16 buttons) and it worked fine so far. I, then, installed version 0.62 of dosbox and my superbox 3 didn't work anymore. So now, I'm back to version 0.61, waiting for a new version working with my superbox 3.

Oh, yeah i'm running Windows XP Pro (5.1 build 2600) with SP2 installed.

Reply 23 of 46, by iampiti

User metadata
Rank Member
Rank
Member
Anonymous wrote:

I've been using dosbox 0.61 with a superbox 3 (playstation pad adapter recognized as a Microsoft Joystick with 4 axes and 16 buttons) and it worked fine so far. I, then, installed version 0.62 of dosbox and my superbox 3 didn't work anymore.

Is your superbox 3 this?
http://www.mayflash.com/pc/pc001/pc001-1.asp.

If it is then it's made by the same company that makes my super dual box. And my super dual box works ok with 0.62.
It's strange. So far I haven't seen anyone else reporting their joystick stopped working in 0.62.

Iampiti

Reply 24 of 46, by GuestCat

User metadata

I am experiencing something really strange with my DOSBox 0.62

Having installed it about a week ago, everything ran fine and absolutely satisfactory. Then today I wanted to open the DOSBox, and a message appeared saying something like "Fatal segmentation fault SDL.DLL"

I have not changed any settings since first editing the .config (mounting drives). The Antivirus program ran before, and nothing ever happened.

As an emergency, I tried to reinstall 0.61 and got the message "SDL can't set palette"

Does anybody have a clue what this could be?

Thanks,

Cat
😖

Reply 28 of 46, by Guest

User metadata

edit: This problem seems to be only, if you resize it from window to fullscreen by ALT+ENTER. If you run DosBox directly in full screen, there is no problem.

And.. why is fullscreen slower then windowed mode? (in dynamic CPU core) I run it in window, and CPU usage is at 70%. I run it in Full Screen and CPU usage is at 95-100% (and sounds are lagging).

I play it at athlon XP 2500+ . CPU cycles are set to 26000. When I am playing doom 2 in window, it goes smoothly. But when I switch to full screen, the CPU goes to 100% and sounds are laging. => Full screen is slower than windowed mode.

It wasn't in 0.61

And... Frameskip doesn't reduce CPU usage level as much as in 0.61. In matter of fact, almost none.

Reply 29 of 46, by jal

User metadata
Rank Oldbie
Rank
Oldbie

Another small bug which may have been reported before, and also exists in previous versions of DOSbox: when switching to the start menu in Windows (Ctrl-Esc) and switching back to Dosbox, it acts like the Esc-key is pressed. Seems like the loss of focus on Ctrl-Esc isn't handeled correctly.

JAL

Reply 30 of 46, by Mephisto1982

User metadata

dosbox 0.62 segfaults on me. i've tried 2 different rpm's from different repositories and they both have the same problem. 0.61 didnt do this, and neither does 0.60. since i couldnt find a 0.61 rpm and dont feel like compiling code now, i went back to 0.60 for now. 0.60, of course, works (like it used to do) but dosbox in general seems to have some very bad code with memory management. This is some output from the fedora core 3 glibc, which has special checks for malloc and free built-in:

[mephisto@chii ~]$ dosbox malloc: using debugging hooks *** glibc detected *** free(): invalid pointer: 0x089768b0 *** *** glibc […]
Show full quote

[mephisto@chii ~]$ dosbox
malloc: using debugging hooks
*** glibc detected *** free(): invalid pointer: 0x089768b0 ***
*** glibc detected *** free(): invalid pointer: 0x08976b30 ***
*** glibc detected *** free(): invalid pointer: 0x089768d0 ***
*** glibc detected *** free(): invalid pointer: 0x08976b60 ***
*** glibc detected *** free(): invalid pointer: 0x089768f0 ***
*** glibc detected *** free(): invalid pointer: 0x08976b90 ***
*** glibc detected *** free(): invalid pointer: 0x08976910 ***
*** glibc detected *** free(): invalid pointer: 0x08976bc0 ***
malloc: using debugging hooks
malloc: using debugging hooks
- using device default
- server format: sample rate = 44100 Hz
- using device default
- server format: sample rate = 44100 Hz
CONFIG:Can't find config file dosbox.conf, using default settings
malloc: using debugging hooks
malloc: using debugging hooks
- using device default
- server format: sample rate = 44100 Hz
- using device default
- server format: sample rate = 44100 Hz
MIDI:Opened device:none
Using joystick Analog 4-axis 4-button joystick with 4 axes and 4 buttons
*** glibc detected *** free(): invalid pointer: 0x089765b8 ***
[mephisto@chii ~]$

0.62 does this:

[mephisto@chii ~]$ dosbox malloc: using debugging hooks *** glibc detected *** free(): invalid pointer: 0x09b2eba8 *** *** glibc […]
Show full quote

[mephisto@chii ~]$ dosbox
malloc: using debugging hooks
*** glibc detected *** free(): invalid pointer: 0x09b2eba8 ***
*** glibc detected *** free(): invalid pointer: 0x09b2ee28 ***
*** glibc detected *** free(): invalid pointer: 0x09b2ebc8 ***
*** glibc detected *** free(): invalid pointer: 0x09b2ee58 ***
*** glibc detected *** free(): invalid pointer: 0x09b2ebe8 ***
*** glibc detected *** free(): invalid pointer: 0x09b2ee88 ***
*** glibc detected *** free(): invalid pointer: 0x09b2ec08 ***
*** glibc detected *** free(): invalid pointer: 0x09b2eeb8 ***
malloc: using debugging hooks
malloc: using debugging hooks
- using device default
- server format: sample rate = 44100 Hz
- using device default
- server format: sample rate = 44100 Hz
CONFIG:Loading settings from config file /home/mephisto/.dosboxrc
Fatal signal: Segmentation Fault (SDL Parachute Deployed)
Killed
[mephisto@chii ~]$

This is from the release notes from FC3, which should give more info about the point i'm trying to make:

* The version of glibc provided with Fedora Core 3 performs additional internal sanity checks to prevent and detect data co […]
Show full quote

* The version of glibc provided with Fedora Core 3 performs additional internal sanity checks to prevent and detect data corruption as early as possible. By default, should corruption be detected, a message similar to the following will be displayed on standard error (or logged via syslog if stderr is not open):

*** glibc detected *** double free or corruption: 0x0937d008 ***

By default, the program that generated this error will also be killed; however, this (and whether or not an error message is generated) can be controlled via the MALLOC_CHECK_ environment variable. The following settings are supported:
o 0 — Do not generate an error message, and do not kill the program
o 1 — Generate an error message, but do not kill the program
o 2 — Do not generate an error message, but kill the program
o 3 — Generate an error message and kill the program
Note

If MALLOC_CHECK_ is explicitly set a value other than 0, this causes glibc to perform more tests that are more extensive than the default, and may impact performance.

Should you have a program from a third party ISV that triggers these corruption checks and displays a message, you should file a defect report with the application's vendor, since this indicates a serious bug.

here is the backtrace from the 0.62 segfault:

0xffffe410 in __kernel_vsyscall () (gdb) bt #0 0xffffe410 in __kernel_vsyscall () #1 0xb7c12955 in raise () from /lib/tls/libc […]
Show full quote

0xffffe410 in __kernel_vsyscall ()
(gdb) bt
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7c12955 in raise () from /lib/tls/libc.so.6
#2 0xb7c14319 in abort () from /lib/tls/libc.so.6
#3 0xb7c4ba1b in malloc_printerr () from /lib/tls/libc.so.6
#4 0xb7c4c4ba in free () from /lib/tls/libc.so.6
#5 0xb6dffe65 in operator delete () from /usr/lib/libstdc++.so.6
#6 0xb6de0fbf in std::string::_Rep::_M_destroy () from /usr/lib/libstdc++.so.6
#7 0xb6f27d61 in Arts::MCOPUtils::createFilePath () from /usr/lib/libmcop.so.1
#8 0xb6f0fb60 in Arts::Dispatcher::Dispatcher () from /usr/lib/libmcop.so.1
#9 0xb73780a4 in arts_backend_init () from /usr/lib/libartscbackend.so.0
#10 0xb73642b6 in arts_init () from /usr/lib/libartsc.so.0
#11 0xb7ea90fa in SDL_MixAudio_MMX_S8 () from /usr/lib/libSDL-1.2.so.0
#12 0xb7ea4b10 in SDL_AudioInit () from /usr/lib/libSDL-1.2.so.0
#13 0xb7ea34fa in SDL_InitSubSystem () from /usr/lib/libSDL-1.2.so.0
#14 0xb7ea35fb in SDL_Init () from /usr/lib/libSDL-1.2.so.0
#15 0x080f4f0a in main ()

i know, there's only system libraries in the backtrace, but since the problem didnt occur with 0.61 and i got 0.60 running here now, i really think its a dosbox bug. I'm running Fedora Core 3 with vanilla kernel 2.6.9. if you need more info on this bug, let me know. i'd also really like to see those free()'s cleaned up, since they prevent dosbox from running on FC3 if i dont set the MALLOC_CHECK_ variable.

Reply 31 of 46, by Qbix

User metadata
Rank DOSBox Author
Rank
DOSBox Author

those malloc errors happen way too early.

I would really think as the backtrace indicates. that arts with sdl is to blame.

Could you disable arts before starting dosbox to see if it changes the error in any way. ?
Because:

#14 0xb7ea35fb in SDL_Init () from /usr/lib/libSDL-1.2.so.0
#15 0x080f4f0a in main ()

Is something like our 3 line of code in dosbox. There have been various reports of arts +sdl crashing on the other hand (many gentoo people reported it. Some Fedora users.)

To give you some insights in the internals of dosbox:

1. init sdl
2. read configfile
3. really start dosbox.

You could try some more sdl games. to see if those free() warnings happen there as well. (Dosbox does some memory allocation when it comes to reading the configfile but not very much before. )

I will take a look at those allocations, but valgrind does like them. (we alloc memory in a certain format. zero it and use it in a different format)

Water flows down the stream
How to ask questions the smart way!

Reply 32 of 46, by Guest

User metadata

i dont know why its outputting stuff about arts. i'm not using arts nor kde. i dont even have any arts deamon running at all.
about the sdl thing: i did have some small games not working anymore (dont know if they use sdl or not), but the majority of the sdl games still work, so i doubt its sdl related. big major sdl games, like Unreal Tournament and Quake 3, work fine.

Reply 35 of 46, by Mephisto1982

User metadata

i have always used the sdl packages provided with the distro. i tried an old version, but this had the same problem. i doubt its arts support, since the old packages have that too and they always worked right before. ive never used arts since i installed the OS.
you were right about UT and Q3 though, i forgot (:blah:), but they were just badly chosen examples. other (non-static) sdl games, like frozen bubble and neverball, work right as well.

Reply 36 of 46, by Titler

User metadata

I believe there is a bug with alt-tabbing and DOSBox; at first I wondered if there was a problem with my front end, but you can see D-Fend's author insist this cannot be the case here;

http://www.mypixels.dk/forums/index.php?showtopic=561

It's not any particular game (I've tried with Ultima Underworlds, Savage Empire, Martian Dreams, Kingmaker, and a plain DOSBox window so far), and alt-tabbing away from the full screen DOSBox window causes crashes 90% of the time... Anyone else have this occur?

Reply 37 of 46, by KillerQ13

User metadata

Crash if the console window is closed. (Missing or broken console handler routine)

Video output problems:

Surface/DDraw (is there a difference?)
Hard crash if the task is switched while DOSBox is in full screen mode. This is extremely annoying with instant messages (as if instant messenges aren't annoying enough already).

DOSBox fails to render anything if the video display settings are changed and will hard crash during a switch to full screen if a secondary monitor is enabled/disabled in that change.

Overlay
Hard crash on video display settings change.

OpenGL/OpenGLNB
If the task is switched while in full screen mode and the task is changed back to DOSBox, the display is offset to the right. If this is repeated, the display continues to offset.

Reply 38 of 46, by Mephisto1982

User metadata

just wanted to let you know that something somehow got fixed with some package updates lately. although 0.60 now segfaults, the FC3 package of 0.62 from the newrpms repository seems to work fine. i really dont know what the problem was but, since everything works now, i doubt if was dosbox itself that was causing the prob. thanks for making this great piece of software anyway 😀