10.4 build attempts?

[JeroMiya]

I was able to build cal3d and CS using the new fink binaries for libjpeg, libpng, libogg, libvorbis, libmng, and libmikmod (not sure if the last two are strictly needed for PS), but cel plugins refused to build. Anybody else get further than I did with xcode 2.1 tools and mac os 10.4?

Oh, and both were massively puking warnings about classes with virtual functions but non virtual destructors.

I heard a bunch of noise on the CS mailing list about gcc 4.x and Tiger compatibility and then nothing. CS builds ok (sans warning overload) so maybe cel needs to be fixed too?

Trying again with gcc_select 3.3 now hopefully that will work.

-Jerm

[backeroo]

Tried this some time before, but didn\'t got CS to compile.
Wich version do you use ( or which snapshot )?
As i didn\'t got CS to work, i couldn\'t try cel, so far.

regards

  backeroo

UPDATE :

just got ps compiled on 10.4.2 with gcc 3.3. I\'ve used the latest CVS-Versions, and now, i have a 0.3.011 Version of Planeshift and the Server will not let me in. :-)
So, i\'ll stick to DFryers binaries for now.

[Kias]

Hey Backeroo, are you able to log on? Just wondering if I should maybe try compiling or wait till Dfryer makes a post.

[backeroo]

Nah, current CVS Version is broken for MacOSX. I\'m also not sure why they didn\'t tag the 011 Release. I didn\'t find a Tag in the Webcvs. So i\'m not sure if i can go back to an older cvs release which worked.
The Problems with the latest cvs version for psclient are in  client/modemanager.cpp and in client/pssoundmngr.cpp.

And after all, the Server still seems to be closed according to http://laanx.fragnetics.com/

regards

  backeroo

EDIT : It\'s also possible that the latest CS/cel cvs versions have broken my build.

[Kias]

Ahh, gotcha...only just noticed the server was offline after I finished downloading the updates.

Oh well

[r.guppy]

pc is now online, hope we dont have to long a wait

[backeroo]

Are there any other Mac-People out there, trying to build 011?
If Yes, does anyone suceeded building the psclient?

[Miya]

Originally posted by backeroo
Are there any other Mac-People out there, trying to build 011?
If Yes, does anyone suceeded building the psclient?

I\'ve gotten CS to build but walktest is crashing both when I run it normally and when I run it with -relight and/or -recalc (two different crashes in different places).

I\'m going to keep going with cel and planeshift but I\'m not so sure that it will work if walktest doesn\'t.

[Miya]

Here are the crashes in walktest:
When walktest is run with -relight option

#0  0x0f5f8064 in csFlags::Get (this=0x0) at ./include/csutil/flags.h:85
#1  0x0f5cdc1c in CastShadows_Front2Back (treenode=0xeaf9470, userdata=0xbfffee9c, cur_timestamp=2, planes_mask=@0xbfffedfc) at plugins/culling/frustvis/frustvis.cpp:1110
#2  0x0f5df688 in csKDTree::Front2Back (this=0xeaf9470, pos=@0xeaf97f4, func=0xf5cd8a0 , userdata=0xbfffee9c, cur_timestamp=2, frustum_mask=0) at libs/csgeom/kdtree.cpp:637
#3  0x0f5dfa5c in csKDTree::Front2Back (this=0xeaf9470, pos=@0xeaf97f4, func=0xf5cd8a0 , userdata=0xbfffee9c, frustum_mask=0) at libs/csgeom/kdtree.cpp:698
#4  0x0f5d0e30 in csFrustumVis::CastShadows (this=0xeaf9e40, fview=0xbffff1f0) at plugins/culling/frustvis/frustvis.cpp:1192
#5  0x049699f8 in csSector::RealCheckFrustum (this=0x9bf8520, lview=0xbffff1f0) at plugins/engine/3d/sector.cpp:826
#6  0x0496a290 in csSector::CheckFrustum (this=0x9bf8520, lview=0xbffff1f0) at plugins/engine/3d/sector.cpp:816
#7  0x0493d0b4 in csLight::CalculateLighting (this=0x9bf9f50) at plugins/engine/3d/light.cpp:376
#8  0x0492feb0 in csEngine::ShineLights (this=0x205d600, region=0x0, meter=0x1b83510) at plugins/engine/3d/engine.cpp:1514
#9  0x04924c20 in csEngine::Prepare (this=0x205d600, meter=0x1b83510) at plugins/engine/3d/engine.cpp:1187
#10 0x0002f760 in WalkTest::Initialize (this=0x2004e00, argc=2, argv=0xbffff88c, iConfigName=0xe9e44 \"/config/walktest.cfg\") at apps/walktest/walktest.cpp:1323
#11 0x00030728 in main (argc=2, argv=0xbffff88c) at apps/walktest/walktest.cpp:1483

Here visobj_wrap->mesh->GetFlags ()  in fustvis.cpp:1110 returns NULL

The second crash occurs when run without any options. Here is the backtrace:
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0xffffffff
#0  0x0f5c0804 in csRef::operator= (this=0xf4de3e0, newobj=0xffffffff) at ./include/csutil/ref.h:234
#1  0x0f596c44 in csFrustumVis::RegisterVisObject (this=0xf4de1c0, visobj=0x21afc20) at plugins/culling/frustvis/frustvis.cpp:256
#2  0x0496a514 in csSector::RegisterEntireMeshToCuller (this=0xeabe430, mesh=0x21afc28) at plugins/engine/3d/sector.cpp:152
#3  0x0496a880 in csSector::SetVisibilityCullerPlugin (this=0xeabe430, plugname=0x4a7c564 \"crystalspace.culling.frustvis\", culler_params=0x0) at plugins/engine/3d/sector.cpp:261
#4  0x0496884c in csSector::GetVisibilityCuller (this=0xeabe430) at plugins/engine/3d/sector.cpp:268
#5  0x04927aec in csEngine::PrecacheDraw (this=0x205d600, region=0x0) at plugins/engine/3d/engine.cpp:1646
#6  0x0002f82c in WalkTest::Initialize (this=0x2004e00, argc=1, argv=0xbffff8a0, iConfigName=0xe9e44 \"/config/walktest.cfg\") at apps/walktest/walktest.cpp:1330
#7  0x00030728 in main (argc=1, argv=0xbffff8a0) at apps/walktest/walktest.cpp:1483

Here mesh->GetShadowCaster ()  in fustvis.cpp:256 returns 0xffffffff, causing csRef::operator= to crash.

[Miya]

In addition cel plugins fail to build. They give errors like this one:

LinkPlugin mgr_quests.csbundle
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (1))
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (3))
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry ()
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (10))
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (17))
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (19))
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (24))
/usr/bin/ld: /Users/jeremybell/PS/CS/./out/macosxppc/debug/libs/libcrystalspace.a(timer.o) malformed object, illegal reference for -dynamic code (reference to a coalesced section (__TEXT,__textcoal_nt) from section (__TEXT,__text) relocation entry (26))
collect2: ld returned 1 exit status

    export MACOSX_DEPLOYMENT_TARGET=10.4
    g++ -bundle -o mgr_quests.csbundle ./out/macosxppc/debug/plugins/tools/quests/quests.o ./out/macosxppc/debug/plugins/tools/quests/reward_changeproperty.o ./out/macosxppc/debug/plugins/tools/quests/reward_debugprint.o ./out/macosxppc/debug/plugins/tools/quests/reward_inventory.o ./out/macosxppc/debug/plugins/tools/quests/reward_newstate.o ./out/macosxppc/debug/plugins/tools/quests/reward_sequence.o ./out/macosxppc/debug/plugins/tools/quests/reward_sequencefinish.o ./out/macosxppc/debug/plugins/tools/quests/seqop_debugprint.o ./out/macosxppc/debug/plugins/tools/quests/seqop_light.o ./out/macosxppc/debug/plugins/tools/quests/seqop_movepath.o ./out/macosxppc/debug/plugins/tools/quests/seqop_transform.o ./out/macosxppc/debug/plugins/tools/quests/trig_entersector.o ./out/macosxppc/debug/plugins/tools/quests/trig_inventory.o ./out/macosxppc/debug/plugins/tools/quests/trig_meshentersector.o ./out/macosxppc/debug/plugins/tools/quests/trig_meshsel.o ./out/macosxppc/debug/plugins/tools/quests/trig_propertychange.o ./out/macosxppc/debug/plugins/tools/quests/trig_sequencefinish.o ./out/macosxppc/debug/plugins/tools/quests/trig_timeout.o ./out/macosxppc/debug/plugins/tools/quests/trig_trigger.o  -g3 -DCS_DEBUG -Wl,-framework,AppKit -Wl,-framework,Foundation -bundle -L/Users/jeremybell/PS/CS/./out/macosxppc/debug/libs -lcrystalspace -lz -lm -lmx -ldl -Wl,-multiply_defined,suppress -L/usr/local/lib -g3 -Wl,-framework,AppKit -Wl,-framework,Foundation -g3 -DCS_DEBUG -Wl,-framework,AppKit -Wl,-framework,Foundation

[edit: note that this fails to load even if MACOSX_DEPLOYMENT_TARGET is the original value]

[haleden]

Originally posted by Miya
Here are the crashes in walktest:
When walktest is run with -relight option

#0  0x0f5f8064 in csFlags::Get (this=0x0) at ./include/csutil/flags.h:85
...
Here visobj_wrap->mesh->GetFlags ()  in fustvis.cpp:1110 returns NULL

The second crash occurs when run without any options. Here is the backtrace:
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0xffffffff
#0  0x0f5c0804 in csRef::operator= (this=0xf4de3e0, newobj=0xffffffff) at ./include/csutil/ref.h:234
...

Here mesh->GetShadowCaster ()  in fustvis.cpp:256 returns 0xffffffff, causing csRef::operator= to crash.

i\'m new to CS, so i don\'t know if this has already been dealt with, but here is my $0.02

i actually got the opposite behavior, at least in that i got the 0xffffffff from GetShadowCaster() when i did the --relight option.

drugeing around with gdb, i found that the reason GetShadowCaster() was returning 0xffffffff is that it was really calling RemoveObj() (or something like that--doing this from memory), which called .find() on its list of Children, which returned csArrayItemNotFound which is basically -1.

the fact that the wrong method is being called smells like a complier problem.

so, i installed the latest XCode preview release with gcc 4.01 and it works!

any other experiences?

hal

[Cha0s]

Interesting, is CEL compiling as well? I\'ve been having the same issues as Miya for a while now...

[haleden]

as i said, i\'m new to CS, so i was trying to get that working first.

i\'ll probably take a crack at CEL this weekend.

hal

[Cha0s]

Great! Tell me how it goes.

[backeroo]

It seems, there is a new release coming up. My last build ( which went fine so far ) comes up with 03.012 in the Title bar. So let\'s get ready to rumble.

regards

  backeroo