0.8.2 and (today's) SVN segfaults with "The Dig" (

Pajarico · Post by **Pajarico** » Tue Mar 14, 2006 10:12 pm

Hi, I have the spanish version of this game. It has english voices and spanish subtitles.

My first question is if this is supported, because in the README (section 5.2) it says scummvm has language options but I don't know if those options are supposed to be official ones and other versions might work but are not officially supported?
Or only ch, ko and jp versions will work?
Or that is for multilanguage editions?

Is there something that keeps a different language not listed there from working?

Supposing my version is good:
Why the segfault then?

Code: Select all

lxuser@hal2000 ~/Compilation/scummvm $ ../bin/scummvm  --debuglevel=2
Debuglevel &#40;from command line&#41;&#58; 2
File default-theme.zip not found
File default-theme.ini not found
File default-theme.zip not found
WARNING&#58; Can not find theme config file 'default-theme.ini'!
WARNING&#58; Could not initialize your preferred theme, falling back to classic style!
File classic.ini not found
Looking for dig
Trying to start game 'The Dig'
File dig.000 not found
File The Dig Data not found
File The Dig Demo Data not found
Output sample rate&#58; 22050 Hz
vars&#91;200&#93; = 0 &#40;via script-0&#41;
vars&#91;200&#93; = 1 &#40;via script-0&#41;
vars&#91;201&#93; = 0 &#40;via script-0&#41;
vars&#91;201&#93; = 2 &#40;via script-0&#41;
vars&#91;228&#93; = 0 &#40;via script-0&#41;
vars&#91;228&#93; = 3 &#40;via script-0&#41;
vars&#91;229&#93; = 0 &#40;via script-0&#41;
vars&#91;229&#93; = 4 &#40;via script-0&#41;
vars&#91;244&#93; = 0 &#40;via script-0&#41;
vars&#91;244&#93; = 5 &#40;via script-0&#41;
vars&#91;254&#93; = 0 &#40;via script-0&#41;
vars&#91;254&#93; = 6 &#40;via script-0&#41;
vars&#91;255&#93; = 0 &#40;via script-0&#41;
vars&#91;255&#93; = 7 &#40;via script-0&#41;
vars&#91;256&#93; = 0 &#40;via script-0&#41;
vars&#91;256&#93; = 8 &#40;via script-0&#41;

<cut>

vars&#91;434&#93; = 0 &#40;via script-0&#41;
vars&#91;434&#93; = 29 &#40;via script-0&#41;
vars&#91;435&#93; = 0 &#40;via script-0&#41;
vars&#91;435&#93; = 30 &#40;via script-0&#41;
vars&#91;436&#93; = 0 &#40;via script-0&#41;
vars&#91;436&#93; = 31 &#40;via script-0&#41;
File dig.sou not found
File dig.sof not found
File dig.sog not found
File dig.so3 not found
File monster.sou not found
File monster.sof not found
File monster.sog not found
File monster.so3 not found
File dig.tlk not found
Adding object 653 from room 93 into inventory
Adding object 40 from room 1 into inventory
Adding object 43 from room 2 into inventory
Adding object 139 from room 17 into inventory
DEBUG&#58; frame-counter 228
DEBUG&#58; frame-counter 138
Adding object 852 from room 93 into inventory
File sq1.trs not found
Fatal signal&#58; Segmentation Fault &#40;SDL Parachute Deployed&#41;

I don't know why all the complains about missing files. There are not such files in my CD and everything in the needed is already copied - I did it twice to discard data corruption -.

The binary is self-compiled in both cases.
gcc version 3.4.4
libsdl-1.2.8

Ask for any missing info.

Thanks in advance.

clem · Post by **clem** » Wed Mar 15, 2006 6:37 am

you generally can't change languages with Lucasarts games in ScummVM (they only come with 1 language in the datafile) - if you have the spanish version it should be detected as spanish (and make sure the settings stay that way) It might be different with the DIG, there's this language.trs file, maybe it'll run in English if you remove that (but I'm just guessing here)

as for the missing files errors, you might get most of them since you run at debuglevel 2; for example:

File dig.sou not found
File dig.sof not found
File dig.sog not found
File dig.so3 not found

means it looks for the uncompressed first (.sou), then for the FLAC compressed (.sof), the .ogg compressed (.sog) and finally the mp3 compressed (.so3) - but it should find at least one of those I think.

it finally stops at not finding sq1.trs, do you have that file? I think that should be in the VIDEO subfolder (I'm sure you know it, but the datafiles are listed at http://scummvm.org/documentation.php?view=datafiles )

Pajarico · Post by **Pajarico** » Wed Mar 15, 2006 7:49 am

clem wrote:you generally can't change languages with Lucasarts games in ScummVM (they only come with 1 language in the datafile) - if you have the spanish version it should be detected as spanish (and make sure the settings stay that way)

The language setting is on "default".

It might be different with the DIG, there's this language.trs file, maybe it'll run in English if you remove that (but I'm just guessing here)

I'm not sure about this one, but I think isn't on the CD either. I will confirm this when I arrive home.

as for the missing files errors, you might get most of them since you run at debuglevel 2; for example:

File dig.sou not found
File dig.sof not found
File dig.sog not found
File dig.so3 not found

means it looks for the uncompressed first (.sou), then for the FLAC compressed (.sof), the .ogg compressed (.sog) and finally the mp3 compressed (.so3) - but it should find at least one of those I think.

I need a confirmation about this. There are no such files in any folder on the CD. I thought that with the DIGVOICE.BUN and DIGMUSIC.BUN would be enough.

it finally stops at not finding sq1.trs, do you have that file? I think that should be in the VIDEO subfolder (I'm sure you know it, but the datafiles are listed at http://scummvm.org/documentation.php?view=datafiles )

That's the list I looked at. The whole VIDEO subfolder was copied and that file doesn't exist. Someone is playing without it?

Post by **eriktorbjorn** » Wed Mar 15, 2006 8:37 am

Pajarico wrote:
it finally stops at not finding sq1.trs, do you have that file? I think that should be in the VIDEO subfolder (I'm sure you know it, but the datafiles are listed at http://scummvm.org/documentation.php?view=datafiles )
That's the list I looked at. The whole VIDEO subfolder was copied and that file doesn't exist. Someone is playing without it?

I don't have it either, but I think that's normal. ScummVM supports so many games that does things in slightly different ways, so often it will try to open a file, notice that it isn't there and decide that "oh, this is one of the games that does things differently".

In this case, I think SmushPlayer::readString() first tries to open the .trs file corresponding to sq1.san. But if that fails, it will try to open digtxt.trs instead.

Vinterstum · Post by **Vinterstum** » Wed Mar 15, 2006 10:38 am

Try debuglevel 9 and use something like gdb or valgrind to determine the exact cause of the crash.

Pajarico · Post by **Pajarico** » Wed Mar 15, 2006 1:47 pm

eriktorbjorn wrote: In this case, I think SmushPlayer::readString() first tries to open the .trs file corresponding to sq1.san. But if that fails, it will try to open digtxt.trs instead.

The inly *.trs file inside VIDEO is digtxt.trs, which are the spanish dialogues, so I guess everything is Ok on that side.

Vinterstum wrote:Try debuglevel 9 and use something like gdb or valgrind to determine the exact cause of the crash.

The log is quite large. Should I post it here? or is there any page that can host logs?

Post by **eriktorbjorn** » Wed Mar 15, 2006 1:58 pm

Pajarico wrote: The log is quite large. Should I post it here? or is there any page that can host logs?

Only the last dozen or so lines are likely to be relevant. (Of course, a backtrace from GDB would probably be even more relevant.)

Pajarico · Post by **Pajarico** » Wed Mar 15, 2006 2:25 pm

--debuglevel=9
I have included more lines because I found strange the lockin/unlocking Mutex in the sense that is the only time it appears while executing. Maybe is just normal, don't know.

Code: Select all

<cut>

Script 71, offset 0x99&#58; &#91;66&#93; o6_stopObjectCode&#40;&#41;
getResourceAddress&#40;Script,74&#41; == 0xb4e404
Script 74, offset 0x56&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 0
Script 74, offset 0x59&#58; &#91;7C&#93; o6_stopScript&#40;&#41;
pop 0
getResourceAddress&#40;Script,5&#41; == 0xb4de44
Script 5, offset 0xd9&#58; &#91;66&#93; o6_stopObjectCode&#40;&#41;
getResourceAddress&#40;Script,1&#41; == 0xb478d4
Script 1, offset 0x176f&#58; &#91;66&#93; o6_stopObjectCode&#40;&#41;
Locking mutex IMuseDigital&#58;&#58;callback&#40;&#41;
Unlocking mutex IMuseDigital&#58;&#58;callback&#40;&#41;
Locking mutex IMuseDigital&#58;&#58;callback&#40;&#41;
Unlocking mutex IMuseDigital&#58;&#58;callback&#40;&#41;
getResourceAddress&#40;Script,13&#41; == 0xa6c174
Script 13, offset 0x1a&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 13, offset 0x1d&#58; &#91;B1&#93; o6_delaySeconds&#40;&#41;
pop 1
getResourceAddress&#40;Room,1&#41; == 0xb96e94
Script 2001, offset 0x1a1d&#58; &#91;6B&#93; o6_cursorCommand&#40;&#41;
Script 2001, offset 0x1a1f&#58; &#91;6B&#93; o6_cursorCommand&#40;&#41;
Script 2001, offset 0x1a21&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 2001, offset 0x1a24&#58; &#91;43&#93; o6_writeWordVar&#40;&#41;
pop 1
writeVar&#40;32774, 1&#41;
Script 2001, offset 0x1a27&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 2001, offset 0x1a2a&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 2001, offset 0x1a2d&#58; &#91;68&#93; o6_cutscene&#40;&#41;
pop 1
pop 1
getResourceAddress&#40;Script,2&#41; == 0xa6c4d4
runScript&#40;Global-2&#41; from 2001-1
getResourceAddress&#40;Script,2&#41; == 0xa6c4d4
Script 2, offset 0x9&#58; &#91;3&#93; o6_pushWordVar&#40;&#41;
readvar&#40;219&#41;
push 0
Script 2, offset 0xc&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 2, offset 0xf&#58; &#91;14&#93; o6_add&#40;&#41;
pop 1
pop 0
push 1
Script 2, offset 0x10&#58; &#91;43&#93; o6_writeWordVar&#40;&#41;
pop 1
writeVar&#40;219, 1&#41;
Script 2, offset 0x13&#58; &#91;6B&#93; o6_cursorCommand&#40;&#41;
Script 2, offset 0x15&#58; &#91;6B&#93; o6_cursorCommand&#40;&#41;
Script 2, offset 0x17&#58; &#91;3&#93; o6_pushWordVar&#40;&#41;
readvar&#40;16384&#41;
push 1
Script 2, offset 0x1a&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 2, offset 0x1d&#58; &#91;E&#93; o6_eq&#40;&#41;
pop 1
pop 1
push 1
Script 2, offset 0x1e&#58; &#91;5D&#93; o6_ifNot&#40;&#41;
pop 1
Script 2, offset 0x21&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 0
Script 2, offset 0x24&#58; &#91;43&#93; o6_writeWordVar&#40;&#41;
pop 0
writeVar&#40;32774, 0&#41;
Script 2, offset 0x27&#58; &#91;B4&#93; o6_printLine&#40;&#41;
Script 2, offset 0x29&#58; &#91;B4&#93; o6_printLine&#40;&#41;
getResourceAddress&#40;Charset,2&#41; == 0xb7ded4
getResourceAddress&#40;Charset,2&#41; == 0xb7ded4
getResourceAddress&#40;Charset,2&#41; == 0xb7ded4
Script 2, offset 0x2d&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 1
Script 2, offset 0x30&#58; &#91;6A&#93; o6_freezeUnfreeze&#40;&#41;
pop 1
Script 2, offset 0x31&#58; &#91;66&#93; o6_stopObjectCode&#40;&#41;
getResourceAddress&#40;Room,1&#41; == 0xb96e94
Script 2001, offset 0x1a2e&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 0
Script 2001, offset 0x1a31&#58; &#91;A4&#93; o6_arrayOps&#40;&#41;
pop 0
readvar&#40;123&#41;
writeVar&#40;123, 0&#41;
writeVar&#40;123, 65&#41;
res.createResource&#40;String,65,16&#41;
Script 2001, offset 0x1a3d&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 6
Script 2001, offset 0x1a40&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 0
Script 2001, offset 0x1a43&#58; &#91;1&#93; o6_pushWord&#40;&#41;
push 2
Script 2001, offset 0x1a46&#58; &#91;C9&#93; o6_kernelSetFunctions&#40;&#41;
pop 2
pop 0
pop 6
getResourceAddress&#40;String,65&#41; == 0xbc6104
getResourceAddress&#40;String,65&#41; == 0xbc6104
getResourceAddress&#40;String,65&#41; == 0xbc6104
trying to read text ressources from sq1.trs
File sq1.trs not found
trying to read text ressources from digtxt.trs
trying to read text ressources from digtxt.trs
Fatal signal&#58; Segmentation Fault &#40;SDL Parachute Deployed&#41;

GDB backtrace

Code: Select all

&#40;gdb&#41; run
Starting program&#58; /home/lxuser/Compilation/bin/scummvm 
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#40;no debugging symbols found&#41;
&#91;Thread debugging using libthread_db enabled&#93;
&#91;New Thread 46912522216864 &#40;LWP 21386&#41;&#93;
&#91;New Thread 1082132832 &#40;LWP 21389&#41;&#93;
WARNING&#58; Can not find theme config file 'default-theme.ini'!
WARNING&#58; Could not initialize your preferred theme, falling back to classic style!
Looking for dig
Trying to start game 'The Dig'
&#91;New Thread 1090525536 &#40;LWP 21391&#41;&#93;

Program received signal SIGSEGV, Segmentation fault.
&#91;Switching to Thread 46912522216864 &#40;LWP 21386&#41;&#93;
0x00000000004dee7a in putchar &#40;&#41;
&#40;gdb&#41; bt
#0  0x00000000004dee7a in putchar &#40;&#41;
#1  0x000000000044aa18 in ?? &#40;&#41;
#2  0x000000000044b2c0 in ?? &#40;&#41;
#3  0x00000000004d91ad in ?? &#40;&#41;
#4  0x00000000004d42f7 in ?? &#40;&#41;
#5  0x00000000004dba1a in ?? &#40;&#41;
#6  0x00000000004dcdee in ?? &#40;&#41;
#7  0x000000000041dd86 in ?? &#40;&#41;
#8  0x000000000041d1d2 in ?? &#40;&#41;
#9  0x0000000000410705 in ?? &#40;&#41;
#10 0x0000000000410bf7 in ?? &#40;&#41;
#11 0x00002aaaabc5f674 in __libc_start_main &#40;&#41; from /lib/tls/libc.so.6
#12 0x0000000000404e9a in ?? &#40;&#41;
#13 0x00007ffffffab048 in ?? &#40;&#41;
#14 0x000000000000001c in ?? &#40;&#41;
#15 0x0000000000000001 in ?? &#40;&#41;
#16 0x00007ffffffac4e4 in ?? &#40;&#41;
#17 0x0000000000000000 in ?? &#40;&#41;
&#40;gdb&#41;

Post by **sev** » Wed Mar 15, 2006 7:18 pm

Pajarico wrote: GDB backtrace

Code: Select all

&#40;gdb&#41; bt
#0  0x00000000004dee7a in putchar &#40;&#41;
#1  0x000000000044aa18 in ?? &#40;&#41;

Please, switch to another thread. For me it is usually thread 3 or 4, that one where you will see ScummEngine calls. (gdb command is 'thread N')

Eugene

Pajarico · Post by **Pajarico** » Wed Mar 15, 2006 7:34 pm

Only threads 2 and 3 exists:

Code: Select all

&#40;gdb&#41; thread 2
&#91;Switching to thread 2 &#40;Thread 1082132832 &#40;LWP 28500&#41;&#41;&#93;#0  0x00002aaaabcf3736 in select &#40;&#41; from /lib/tls/libc.so.6
&#40;gdb&#41; bt
#0  0x00002aaaabcf3736 in select &#40;&#41; from /lib/tls/libc.so.6
#1  0x00002aaaab594b06 in SDL_Delay &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#2  0x00002aaaab594b5d in SDL_Delay &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#3  0x00002aaaab593b77 in SDL_RunThread &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#4  0x00002aaaab593e39 in SDL_KillThread &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#5  0x00002aaaab6b20fa in start_thread &#40;&#41; from /lib/tls/libpthread.so.0
#6  0x00002aaaabcf9a73 in clone &#40;&#41; from /lib/tls/libc.so.6
#7  0x0000000000000000 in ?? &#40;&#41;

Code: Select all

&#40;gdb&#41; thread 3
&#91;Switching to thread 3 &#40;Thread 1090525536 &#40;LWP 28503&#41;&#41;&#93;#0  0x00002aaaabcf2e29 in ioctl &#40;&#41; from /lib/tls/libc.so.6
&#40;gdb&#41; bt
#0  0x00002aaaabcf2e29 in ioctl &#40;&#41; from /lib/tls/libc.so.6
#1  0x00002aaaab1a0594 in snd_pcm_hw_fast_tstamp &#40;&#41; from /usr/lib/libasound.so.2
#2  0x00002aaaab560f09 in SDL_FreeWAV &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#3  0x00002aaaab55ca6c in SDL_RunAudio &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#4  0x00002aaaab593b77 in SDL_RunThread &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#5  0x00002aaaab593e39 in SDL_KillThread &#40;&#41; from /usr/lib/libSDL-1.2.so.0
#6  0x00002aaaab6b20fa in start_thread &#40;&#41; from /lib/tls/libpthread.so.0
#7  0x00002aaaabcf9a73 in clone &#40;&#41; from /lib/tls/libc.so.6
#8  0x0000000000000000 in ?? &#40;&#41;

I haven't found those ScumEngine calls:

Code: Select all

&#40;gdb&#41; thread 4
Thread ID 4 not known.

And so on till thread 18 (last i tried).

fingolfin · Post by **fingolfin** » Wed Mar 15, 2006 8:44 pm

We are looking at a stripped release binary here, so it's not big surprise that little to no info in the backtrace is present...

Maybe Pajarico can compile his own version of ScummVM, which includes the debug info by default?

Pajarico · Post by **Pajarico** » Wed Mar 15, 2006 9:10 pm

fingolfin wrote:We are looking at a stripped release binary here, so it's not big surprise that little to no info in the backtrace is present...

Maybe Pajarico can compile his own version of ScummVM, which includes the debug info by default?

Every GDB report I posted is from the SVN, compiled by myself

.
I looked at ./configure --help to see if there was an option to activate debug but turned out that the only option about debugging was to explicitly deactivate it, so the binary should be debuggable in its current form.

I haven't stripped the binary with any elf util either.

What's next?

Pajarico · Post by **Pajarico** » Sun Mar 19, 2006 1:03 pm

Finally I managed to get a debuggable binary, here is the backtrace with those ScummEngine calls:

Code: Select all

&#40;gdb&#41; run
Starting program&#58; /usr/games/bin/scummvm 
&#91;Thread debugging using libthread_db enabled&#93;
&#91;New Thread 46912516417536 &#40;LWP 14567&#41;&#93;
&#91;New Thread 1082132832 &#40;LWP 14570&#41;&#93;
Looking for dig
Trying to start game 'The Dig'
&#91;New Thread 1090525536 &#40;LWP 14572&#41;&#93;

Program received signal SIGSEGV, Segmentation fault.
&#91;Switching to Thread 46912516417536 &#40;LWP 14567&#41;&#93;
Scumm&#58;&#58;NutRenderer&#58;&#58;loadFont &#40;this=0xb008c0, filename=0x7fffffca2be0 "font0.nut"&#41; at endian.h&#58;146
146     endian.h&#58; No such file or directory.
        in endian.h
&#40;gdb&#41; bt
#0  Scumm&#58;&#58;NutRenderer&#58;&#58;loadFont &#40;this=0xb008c0, filename=0x7fffffca2be0 "font0.nut"&#41; at endian.h&#58;146
#1  0x000000000044b392 in Scumm&#58;&#58;SmushPlayer&#58;&#58;setupAnim &#40;this=0xade770, file=0xb0486e "\002"&#41; at smush_player.cpp&#58;1055
#2  0x000000000044bb74 in Scumm&#58;&#58;SmushPlayer&#58;&#58;play &#40;this=0xade770, filename=0xa4cc3a "sq1.san", offset=0, startFrame=0&#41; at smush_player.cpp&#58;1299
#3  0x00000000004d3b2d in Scumm&#58;&#58;ScummEngine_v6&#58;&#58;o6_kernelSetFunctions &#40;this=0x9b32c0&#41; at script_v6.cpp&#58;2524
#4  0x00000000004cecc7 in Scumm&#58;&#58;ScummEngine_v6&#58;&#58;executeOpcode &#40;this=0xac5c80, i=110 'n'&#41; at script_v6.cpp&#58;377
#5  0x00000000004c2fb8 in Scumm&#58;&#58;ScummEngine&#58;&#58;executeScript &#40;this=0x9b32c0&#41; at script.cpp&#58;466
#6  0x00000000004c430d in Scumm&#58;&#58;ScummEngine&#58;&#58;runAllScripts &#40;this=0x9b32c0&#41; at script.cpp&#58;905
#7  0x00000000004187d4 in Scumm&#58;&#58;ScummEngine&#58;&#58;scummLoop &#40;this=0x9b32c0, delta=0&#41; at scumm.cpp&#58;2503
#8  0x0000000000417c02 in Scumm&#58;&#58;ScummEngine&#58;&#58;go &#40;this=0x9b32c0&#41; at scumm.cpp&#58;2271
#9  0x000000000040ec50 in runGame &#40;detector=@0x7fffffca3610, system=@0x95bce0&#41; at main.cpp&#58;310
#10 0x000000000040efc5 in main &#40;argc=1, argv=0x7fffffca3778&#41; at main.cpp&#58;456

I'm running amd64 in case that makes a difference.

Regards.

DeiF · Post by **DeiF** » Thu Apr 12, 2007 7:40 pm

I'm using ScummVM 0.9.1 (AMD64, SuSE 10.2), The Dig spanish version.
ScummVM segfaults when trying to run The Dig.

Code: Select all

Opening hashed&#58; /home/deif/Juegos/DIG/VIDEO/SQ1.SAN
trying to read text resources from sq1.trs
File sq1.trs not found
trying to read text resources from digtxt.trs
Opening hashed&#58; /home/deif/Juegos/DIG/VIDEO/DIGTXT.TRS
trying to read text resources from digtxt.trs
Opening hashed&#58; /home/deif/Juegos/DIG/VIDEO/DIGTXT.TRS
Opening hashed&#58; /home/deif/Juegos/DIG/VIDEO/FONT0.NUT
Violación de segmento

Any ideas?

[EDIT] I deleted FONT0.NUT and copyed FONT3.NUT to FONT0.NUT and then the game starts.
Corrupted file? The original FONT0.NUT has the following md5sum: 32278079230f6797190efdbf759e742e
[EDIT2] Yes, it was file corruption. Sorry for the post. However it shouldn't segfault.