basilisk-devel — Basilisk II development-related discussion

Re: [B2-devel] Re: SheepShaver 2.2 snapshot 2004/07/07

[Gwenole B. <gb...@di...>]

Hi,

>> else if (strcmp(value, "750CX") =3D=3D 0)
>> 	PVR =3D 0x00083214;
>
> OK, I will merge with Linux cputable[].

Done, committed to CVS. There are even patterns for some newer POWER=20
processors, but I have never approached IBM systems with that.

> I have an oldish standalone test program but I have to reboot into=20
> Linux first. ;-)

What can you get out of:
<http://gwenole.beauchesne.free.fr/sheepshaver/paranoia.cpp> ?

Thanks,
Gwenol=E9.

Re: [B2-devel] Ethernet/Internet Networking?

[Gwenole B. <gb...@di...>]

Hi,

> How does one use tun/tap to get SheepShaver to connect to the net?
> I've spent 3 days on it and can't figure it out at all.  Sorry to be a
> bother =/.

First, if you build SheepShaver from sources yourself on a non big 
endian system (e.g. x86), ethernet support is not enabled by default, 
you will have to change kpx_cpu/sheepshaver_glue.cpp for that.

Second, I got the TUN virtual device working only at home through a DSL 
connection (ethernet modem).

Make sure you have the following prefs items set up correctly in 
~/.sheepshaver_prefs:
ether tun
etherconfig /path/to/tunconfig

Make sure you have sudo working for your user for the following 
binaries:
/sbin/ifconfig
/sbin/iptables

Or you will have to run SheepShaver as root or run tunconfig through a 
wrapper script that calls kdesu or whatever.

Next, in the MacOS side, set up a static IP in this form: 172.20.XX.2
Where XX is generally 40, you can check that with /sbin/ifconfig for 
the IP of the tun0 device. Use another IP in that network.

Re: [B2-devel] Re: SheepShaver 2.2 snapshot 2004/07/07

[Gwenole B. <gb...@di...>]

Hi,

> Running Linux/PPC with a 2.4.22 kernel, I got it to compile and run=20
> with a
> small problem, namely, the config script apparently didn't detect that=20=

> my
> system has dirent.h and sys/dir.h, so I had to add "-DHAVE_DIRENT_H
> -DHAVE_SYS_DIR_H" to the CXXFLAGS line in the Makefile or else it=20
> errors out.

Do you still have the config.log file? If so, make a check for=20
"dirent.h" and see what error could occur.

> else if (strcmp(value, "750CX") =3D=3D 0)
> 	PVR =3D 0x00083214;

OK, I will merge with Linux cputable[].

> However, the show-stopper happens after starting the paranoia check,=20=

> which
> causes the program to hang.  A bit of investigation shows that it=20
> never gets
> past the line "set_r2(0xaffebad5);" in Linux/paranoia.cpp.  This=20
> happens
> whether or not I've added those two 750CXe lines to main_unix.cpp.

I have an oldish standalone test program but I have to reboot into=20
Linux first. ;-)
Besides, what are your configure options [summary] ?

Could you tell me more about your system? I am using Mandrakelinux 9.1=20=

with a kernel 2.4.21 + glibc 2.3.1. Can you send the output of running=20=

/lib/libc.so.6 ?

Bye,
Gwenol=E9.=

[B2-devel] Ethernet/Internet Networking?

[Erik A. <ea...@gm...>]

How does one use tun/tap to get SheepShaver to connect to the net? 
I've spent 3 days on it and can't figure it out at all.  Sorry to be a
bother =/.

-Erik Arnson

[B2-devel] Re: SheepShaver 2.2 snapshot 2004/07/07

[Eric H. <eah...@lo...>]

Hi, I'm new to SheepShaver but I thought I'd give this a shot.

Running Linux/PPC with a 2.4.22 kernel, I got it to compile and run with a 
small problem, namely, the config script apparently didn't detect that my 
system has dirent.h and sys/dir.h, so I had to add "-DHAVE_DIRENT_H 
-DHAVE_SYS_DIR_H" to the CXXFLAGS line in the Makefile or else it errors out.

When running, it doesn't detect my 750CXe processor and assumes it's a 604, 
which (after doing a quick search on the web) can be fixed, possibly (keeping 
in mind that I don't really know what I'm doing here), by adding these lines 
to the appropriate spot in main_unix.cpp:

else if (strcmp(value, "750CX") == 0)
	PVR = 0x00083214;

However, the show-stopper happens after starting the paranoia check, which 
causes the program to hang.  A bit of investigation shows that it never gets 
past the line "set_r2(0xaffebad5);" in Linux/paranoia.cpp.  This happens 
whether or not I've added those two 750CXe lines to main_unix.cpp.

Any ideas?  I've done some poking around in the source but I'm afraid I'm 
rather clueless as to what's supposed to happen there....

--Eric

Re: [B2-devel] Native Resource Manager problem

[Gwenole B. <gb...@di...>]

Hi,

> Did you thinked looking at MacOnLinux source?
> I think it maybe be there.

MOL is a too low-level runtime environment. SheepShaver provides=20
high-level reimplementation of certain functions to make it faster, but=20=

of course less compatible.

> If not another solution can be use a debugger on a real macintosh when
> loading macos, but this is more complex :s

Nope, I am working on it, so I'd prefer not corrupt my data. ;-)

I thought Christian would have one of his usual clever answers that=20
would make the fix straightforward. ;-)

Bye,
Gwenol=E9.=

[B2-devel] SheepShaver 2.2 snapshot 2004/07/07

[Gwenole B. <gb...@di...>]

Hi,

I have placed new source tarballs and RPMs at:
<http://gwenole.beauchesne.free.fr/sheepshaver/files/>

Files of interest:
- SheepShaver-2.2-14.src.rpm				source RPM =
package
- SheepShaver-2.2-14.i586.rpm			x86 binary RPM package
- SheepShaver-2.2-20040707.tar.bz2		CVS sources as of this =
morning
- SheepShaver.macosx.20040707.tar.bz2	MacOS X binary

[I will try to make a proper webpage later. ;-)]

User visible changes since 2004/06/09:
- Add support for MacOS 9.0.4
- Native MacOS X port through SDL graphics
- Fix timebase emulation, i.e. fix applications using UpTime()
- Fix hardware cursor acceleration on X11, e.g. make the I-beam cursor=20=

visible
- Fix 'r' and <ctrl> key mappings for Apple X11 servers
- Fix video thread cancellation on MacOS X
- Fix detection of PowerPC 750FX and 970 processors

Bye,
Gwenol=E9.=

Re: [B2-devel] Error Compiling in 0.9... in Solaris 9

[Nigel P. <ni...@in...>]

> But, when I started compiling it, I got the following error:
...
> Undefined first referenced
> symbol in file
> __gxx_personality_v0 obj/readcpu.o
...
> What is this gxx_personality thingie it is looking for?

	I think I had the same error with my OS X port once.
The fix was to use CXX instead of CC (i.e. G++ instead of gcc)
for anything that used readcpu.o (which is built from C++ source).

--
Nigel Pearson, ni...@in...|  "Let's see how Spike is going"
Telstra BI&D, Sydney, Australia      |              ...
Office: 8255 4222    Fax:  8255 3153 |"I'd like to keep Spike as my 
pet."
Mobile: 0408 664435  Home: 9792 6998 |        Illyria - Angel

Re: [B2-devel] Error Compiling in 0.9... in Solaris 9

[Mauricio <ra...@ku...>]

>Hi,
>
>>I am trying to compile basilik 0.9 in this Sun U30 running Solaris 
>>9. Besides the fact it completely failed to see my gtk installation,
>
>Probably it's not in a standard search path?

	Well, it is installed in /usr/local (I am using the package 
off www.sunfreeware.com), as in

/usr/local/lib/libgtk-x11-2.0.la

>Or gtk-config was not found, etc.

	That could be a thought.  I wonder where the bloody thing is hiding...

>>Enable video on SEGV signals ..... : no
>>Addressing mode .................. : memory banks
>
>I'd be curious to know why it failed, considering I used to build on 
>an Ultra 5 at school a few years ago. Any hints from config.cache?

	Even though it is 80-odd lines, I am including config.cache 
in the end of this message.  But, the online video-related line I see 
is this one:

ac_cv_lib_Xxf86vm_XF86VidModeQueryExtension=${ac_cv_lib_Xxf86vm_XF86VidModeQueryExtension='no'}

which seems to expect to be running in a PC.  Am I missing something here?

>>c++ -I../include -I. -I../uae_cpu -DHAVE_CONFIG_H -DOS_solaris 
>>-DCPU_sparc -DBSD_COMP -D_POSIX_PTHREAD_SEMANTICS -DFPU_UAE 
>>-D_REENTRANT -DDATADIR=\"/usr/local/share/BasiliskII\" -O2 
>>-I/usr/openwin/include -c cpudefs.cpp -o obj/cpudefs.o
>>gcc -o obj/gencpu obj/gencpu.o obj/readcpu.o obj/cpudefs.o
>
>Use g++ to link C++ objects built with g++. Or use current CVS. 
>There are source snapshots at Christian's site, but you can (I 
>prefer) fetch current CVS.

	I will try a current CVS and update on my findings then. =)


ac_cv_c_bigendian=${ac_cv_c_bigendian='yes'}
ac_cv_c_const=${ac_cv_c_const='no'}
ac_cv_c_inline=${ac_cv_c_inline='no'}
ac_cv_can_map_lm=${ac_cv_can_map_lm='yes'}
ac_cv_func_atanh=${ac_cv_func_atanh='no'}
ac_cv_func_cfmakeraw=${ac_cv_func_cfmakeraw='no'}
ac_cv_func_clock_gettime=${ac_cv_func_clock_gettime='no'}
ac_cv_func_connect=${ac_cv_func_connect='no'}
ac_cv_func_gethostbyname=${ac_cv_func_gethostbyname='no'}
ac_cv_func_isinf=${ac_cv_func_isinf='no'}
ac_cv_func_isinfl=${ac_cv_func_isinfl='no'}
ac_cv_func_isnan=${ac_cv_func_isnan='no'}
ac_cv_func_isnanl=${ac_cv_func_isnanl='no'}
ac_cv_func_pthread_cancel=${ac_cv_func_pthread_cancel='no'}
ac_cv_func_remove=${ac_cv_func_remove='no'}
ac_cv_func_sem_init=${ac_cv_func_sem_init='no'}
ac_cv_func_shmat=${ac_cv_func_shmat='no'}
ac_cv_func_sigaction=${ac_cv_func_sigaction='no'}
ac_cv_func_signal=${ac_cv_func_signal='no'}
ac_cv_func_strdup=${ac_cv_func_strdup='no'}
ac_cv_func_timer_create=${ac_cv_func_timer_create='no'}
ac_cv_have_extended_signals=${ac_cv_have_extended_signals='no'}
ac_cv_have_sigcontext_hack=${ac_cv_have_sigcontext_hack='no'}
ac_cv_have_x=${ac_cv_have_x='have_x=yes 
ac_x_includes= ac_x_libraries=/usr/lib'}
ac_cv_header_fcntl_h=${ac_cv_header_fcntl_h='yes'}
ac_cv_header_stdc=${ac_cv_header_stdc='no'}
ac_cv_header_sys_time_h=${ac_cv_header_sys_time_h='yes'}
ac_cv_header_time=${ac_cv_header_time='yes'}
ac_cv_header_unistd_h=${ac_cv_header_unistd_h='yes'}
ac_cv_lib_ICE_IceConnectionNumber=${ac_cv_lib_ICE_IceConnectionNumber='no'}
ac_cv_lib_PTL_pthread_create=${ac_cv_lib_PTL_pthread_create='no'}
ac_cv_lib_Xxf86dga_XF86DGAQueryExtension=${ac_cv_lib_Xxf86dga_XF86DGAQueryExtension='no'}
ac_cv_lib_Xxf86vm_XF86VidModeQueryExtension=${ac_cv_lib_Xxf86vm_XF86VidModeQueryExtension='no'}
ac_cv_lib_c_r_pthread_create=${ac_cv_lib_c_r_pthread_create='no'}
ac_cv_lib_dnet_dnet_ntoa=${ac_cv_lib_dnet_dnet_ntoa='no'}
ac_cv_lib_dnet_stub_dnet_ntoa=${ac_cv_lib_dnet_stub_dnet_ntoa='no'}
ac_cv_lib_ipc_shmat=${ac_cv_lib_ipc_shmat='no'}
ac_cv_lib_nsl_gethostbyname=${ac_cv_lib_nsl_gethostbyname='no'}
ac_cv_lib_posix4_sem_init=${ac_cv_lib_posix4_sem_init='no'}
ac_cv_lib_posix_remove=${ac_cv_lib_posix_remove='no'}
ac_cv_lib_pthread_pthread_create=${ac_cv_lib_pthread_pthread_create='no'}
ac_cv_lib_socket_connect=${ac_cv_lib_socket_connect='no'}
ac_cv_path_ESD_CONFIG=${ac_cv_path_ESD_CONFIG='/usr/bin/esd-config'}
ac_cv_path_GTK_CONFIG=${ac_cv_path_GTK_CONFIG='no'}
ac_cv_prog_CC=${ac_cv_prog_CC='g++'}
ac_cv_prog_CPP=${ac_cv_prog_CPP='g++ -E'}
ac_cv_prog_CXX=${ac_cv_prog_CXX='g++'}
ac_cv_prog_cc_cross=${ac_cv_prog_cc_cross='no'}
ac_cv_prog_cc_g=${ac_cv_prog_cc_g='yes'}
ac_cv_prog_cc_g___c_o=${ac_cv_prog_cc_g___c_o='no'}
ac_cv_prog_cc_works=${ac_cv_prog_cc_works='yes'}
ac_cv_prog_cxx_cross=${ac_cv_prog_cxx_cross='no'}
ac_cv_prog_cxx_g=${ac_cv_prog_cxx_g='yes'}
ac_cv_prog_cxx_works=${ac_cv_prog_cxx_works='yes'}
ac_cv_prog_gcc=${ac_cv_prog_gcc='yes'}
ac_cv_prog_gxx=${ac_cv_prog_gxx='yes'}
ac_cv_prog_make_make_set=${ac_cv_prog_make_make_set='yes'}
ac_cv_sigaction_need_reinstall=${ac_cv_sigaction_need_reinstall='no'}
ac_cv_signal_need_reinstall=${ac_cv_signal_need_reinstall='no'}
ac_cv_sizeof_int=${ac_cv_sizeof_int='0'}
ac_cv_sizeof_long=${ac_cv_sizeof_long='0'}
ac_cv_sizeof_long_long=${ac_cv_sizeof_long_long='0'}
ac_cv_sizeof_short=${ac_cv_sizeof_short='0'}
ac_cv_sizeof_void_p=${ac_cv_sizeof_void_p='0'}
ac_cv_struct_tm=${ac_cv_struct_tm='time.h'}
ac_cv_type_loff_t=${ac_cv_type_loff_t='no'}
ac_cv_type_off_t=${ac_cv_type_off_t='yes'}
ac_cv_type_signal=${ac_cv_type_signal='void'}
ac_cv_type_size_t=${ac_cv_type_size_t='yes'}

Re: [B2-devel] Error Compiling in 0.9... in Solaris 9

[Gwenole B. <gb...@di...>]

Hi,

> I am trying to compile basilik 0.9 in this Sun U30 running Solaris 9.=20=

> Besides the fact it completely failed to see my gtk installation,

Probably it's not in a standard search path? Or gtk-config was not=20
found, etc.

> Enable video on SEGV signals ..... : no
> Addressing mode .................. : memory banks

I'd be curious to know why it failed, considering I used to build on an=20=

Ultra 5 at school a few years ago. Any hints from config.cache?

> c++ -I../include -I. -I../uae_cpu -DHAVE_CONFIG_H -DOS_solaris=20
> -DCPU_sparc -DBSD_COMP -D_POSIX_PTHREAD_SEMANTICS -DFPU_UAE=20
> -D_REENTRANT -DDATADIR=3D\"/usr/local/share/BasiliskII\" -O2=20
> -I/usr/openwin/include -c cpudefs.cpp -o obj/cpudefs.o
> gcc -o obj/gencpu obj/gencpu.o obj/readcpu.o obj/cpudefs.o

Use g++ to link C++ objects built with g++. Or use current CVS. There=20
are source snapshots at Christian's site, but you can (I prefer) fetch=20=

current CVS.

Bye,
Gwenol=E9.=

[B2-devel] Error Compiling in 0.9... in Solaris 9

[Mauricio <ra...@ku...>]

I am trying to compile basilik 0.9 in this Sun U30 running Solaris 9. 
Besides the fact it completely failed to see my gtk installation, the 
configuration step went ok:

Basilisk II configuration summary:

XFree86 DGA support .............. : no
XFree86 VidMode support .......... : no
fbdev DGA support ................ : yes
Enable video on SEGV signals ..... : no
ESD sound support ................ : yes
GTK user interface ............... : no
mon debugger support ............. : no
Running m68k code natively ....... : no
Floating-Point emulation core .... : original uae core
Assembly optimizations ........... : none
Addressing mode .................. : memory banks

But, when I started compiling it, I got the following error:

mauricio@poisonivy-18>make
make: Warning: File `Makefile' has modification time 4.5e+02 s in the future
c++ -I../include -I. -I../uae_cpu -DHAVE_CONFIG_H -DOS_solaris 
-DCPU_sparc -DBSD_COMP -D_POSIX_PTHREAD_SEMANTICS -DFPU_UAE 
-D_REENTRANT -DDATADIR=\"/usr/local/
[...]
gcc -o obj/build68k obj/build68k.o
obj/build68k <../uae_cpu/table68k >cpudefs.cpp
c++ -I../include -I. -I../uae_cpu -DHAVE_CONFIG_H -DOS_solaris 
-DCPU_sparc -DBSD_COMP -D_POSIX_PTHREAD_SEMANTICS -DFPU_UAE 
-D_REENTRANT -DDATADIR=\"/usr/local/share/BasiliskII\" -O2 
-I/usr/openwin/include -c cpudefs.cpp -o obj/cpudefs.o
gcc -o obj/gencpu obj/gencpu.o obj/readcpu.o obj/cpudefs.o
Undefined first referenced
symbol in file
__gxx_personality_v0 obj/readcpu.o
ld: fatal: Symbol referencing errors. No output written to obj/gencpu
collect2: ld returned 1 exit status
make: *** [obj/gencpu] Error 1
mauricio@poisonivy-19>pwd
/usr/users/mauricio/todo/BasiliskII-0.9/src/Unix
mauricio@poisonivy-20>

What is this gxx_personality thingie it is looking for?  If I am not 
making the right question, I would appreciate any pointers. =)

RE: [B2-devel] Native Resource Manager problem

[Natalia P. <cl...@te...>]

Did you thinked looking at MacOnLinux source?
I think it maybe be there.

If not another solution can be use a debugger on a real macintosh when
loading macos, but this is more complex :s=20

> -----Mensaje original-----
> De: bas...@li...=20
> [mailto:bas...@li...] En nombre=20
> de Gwenole Beauchesne
> Enviado el: lunes, 05 de julio de 2004 6:29
> Para: bas...@li...
> Asunto: [B2-devel] Native Resource Manager problem
>=20
> Hi,
>=20
> MacOS 9.1 is not booting, neither does MacOS 8.6 when the=20
> machine ID is set to 406 (NewWorld). It seems to be due to=20
> the fact that the Native Resource Manager is not initialized=20
> somehow. i.e. the descriptors at 0x1480 et al. are ROM=20
> locations and MacOS will issue things like "unimplemented=20
> trap" later on.
>=20
> Do you know how the native resource manager could be forced=20
> loaded or what else does MacOS check for to load it?
>=20
> Thanks,
> Gwenol=E9.
>=20
>=20
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 -=20
> digital self defense, top technical experts, no vendor=20
> pitches, unmatched networking opportunities. Visit=20
> www.blackhat.com _______________________________________________
> basilisk-devel mailing list
> bas...@li...
> https://lists.sourceforge.net/lists/listinfo/basilisk-devel

[B2-devel] Looking for G5 users

[Gwenole B. <gb...@di...>]

Hi,

If there are G5 users out there, please test current SheepShaver CVS=20
there. I think I fixed remaining bits for it to be supported but I=20
can't test.

Thanks,
Gwenol=E9.=

[B2-devel] Native Resource Manager problem

[Gwenole B. <gb...@di...>]

Hi,

MacOS 9.1 is not booting, neither does MacOS 8.6 when the machine ID is=20=

set to 406 (NewWorld). It seems to be due to the fact that the Native=20
Resource Manager is not initialized somehow. i.e. the descriptors at=20
0x1480 et al. are ROM locations and MacOS will issue things like=20
"unimplemented trap" later on.

Do you know how the native resource manager could be forced loaded or=20
what else does MacOS check for to load it?

Thanks,
Gwenol=E9.=

Re: [B2-devel] BasiliskII SDL on IRIX

[Gwenole B. <gb...@di...>]

Hi,

> I gave the new SDL support a spin on IRIX.  I found I needed the
> following patches for building:

OK, thanks I indeed can't make a cast value an lvalue.

> The gamma seems to be a bit different than with the X
> driver, but that's OK.

Probably because it's not working? ;-) I really don't know how to make=20=

gamma stuff work with SDL. They do have helper functions but not=20
clearly interfacable means with existing B2/SS code.

> "Thousands" of colors mode squashes the entire
> display into the left half of the window, with an odd green cast.
> Looks like something's not quite right in the 16bpp blitter...

Possible, blitters marked as "NT" (or "OK" in one way only) in=20
Screen_blitters[] are not tested. I normally tested all other big=20
endian blitters when I was still accessing an UltraSPARC station and=20
nowadays on my PBG4. And I tried to be very careful when writing the=20
rest.

On the other hand, if it is only a half of the window, that's a bit=20
odd. I would expect a bug to be consistently propagated on both sides.

> Gdk-ERROR **: BadMatch (invalid parameter attributes)
>   serial 41 error_code 8 request_code 129 minor_code 3

Could be a visual mismatch, it's really picky. Though it should have=20
worked better with consistent use of find_visual_for_depth() when this=20=

was implemented. :-(

BTW, I will likely implement soon a new JIT engine now that I am sure=20
OS 9.0 could work in real addressing mode. The current JIT definitely=20
turned out to be (i) too gcc specific, (ii) architecture dependent=20
anyway. And keeping in mind the MIPS ISA definitely helps in=20
determining a decent way to emulate PPC condition codes. Problem is=20
without proper hardware support, overflow condition would be=20
inefficitently handled.

Thanks,
Gwenol=E9.=

[B2-devel] BasiliskII SDL on IRIX

[Brian J. <bjj...@us...>]

Gwenole,

I gave the new SDL support a spin on IRIX.  I found I needed the
following patches for building:

Index: video_blit.h
===================================================================
RCS file: /cvs/BasiliskII/src/Unix/video_blit.h,v
retrieving revision 1.10
diff -u -r1.10 video_blit.h
--- video_blit.h        26 Jun 2004 17:20:35 -0000      1.10
+++ video_blit.h        29 Jun 2004 20:01:25 -0000
@@ -69,7 +69,7 @@
 #define VIDEO_MODE_X                   mode.x
 #define VIDEO_MODE_Y                   mode.y
 #define VIDEO_MODE_RESOLUTION  mode.resolution_id
-#define VIDEO_MODE_DEPTH               (int)mode.depth
+#define VIDEO_MODE_DEPTH               mode.depth
 #endif

 #endif /* VIDEO_BLIT_H */
Index: ../SDL/video_sdl.cpp
===================================================================
RCS file: /cvs/BasiliskII/src/SDL/video_sdl.cpp,v
retrieving revision 1.9
diff -u -r1.9 video_sdl.cpp
--- ../SDL/video_sdl.cpp        27 Jun 2004 17:31:21 -0000      1.9
+++ ../SDL/video_sdl.cpp        29 Jun 2004 20:01:25 -0000
@@ -413,7 +413,7 @@
        VIDEO_MODE_Y = height;
        VIDEO_MODE_RESOLUTION = resolution_id;
        VIDEO_MODE_ROW_BYTES = bytes_per_row;
-       VIDEO_MODE_DEPTH = depth;
+       VIDEO_MODE_DEPTH = (enum video_depth)depth;
        VideoModes.push_back(mode);
 }

I believe casts in lvalues are a GCC extension -- MIPSPro CC doesn't
like them.

The app configured and built OK.  "Millions" of colors mode seems to
work fine (although it's a bit slower than X11/XShm, it's still very
usable.)  The gamma seems to be a bit different than with the X
driver, but that's OK.  "Thousands" of colors mode squashes the entire
display into the left half of the window, with an odd green cast.
Looks like something's not quite right in the 16bpp blitter...  Other
video modes produce an error like this, just after remapping the
window:

Gdk-ERROR **: BadMatch (invalid parameter attributes)
  serial 41 error_code 8 request_code 129 minor_code 3

Those used to occur quite a bit when we were debugging BasiliskII's
runtime depth switching.  I believe they stemmed from trying to use
the wrong X11 visual...  Unlike XFree86, Xsgi supports a full range of
visuals at different bit depths, all at once.  Some code had been
assuming a single visual, if I recall....

Anyway, keep up the good work.

Thanks,

						Brian Johnson

--------------------------------------------------------------------

   On learning: "The brighter grows the candle flame, the larger the
   diameter of the surrounding darkness"
                                           -- Robyn McNamara

[B2-devel] Status for SheepShaver on MacOS X

[Gwenole B. <gb...@di...>]

Hi,

Current CVS now has correct keyboard support through SDL native=20
scancodes. i.e. you have to use new BasiliskII/src/SDL/keycodes file=20
and set "keycodes" to "true" in your prefs file.

Besides, I added support for the scrap (aka copy-paste) + cdrom code=20
from Basilisk II native OS X port from Nigel. Note that for the latter=20=

to be effective, you have to insert a CD prior to starting the=20
application so that the device is created.

TODO:
- Audio support through SDL/audio, mixer
- Faster video graphics. For that, if someone knows how to setup=20
alternate stacks for Mach exception handlers, please tell me.=20
sigaltstack() won't work as it is intended for (slow) BSD-style signals.

Bye,
Gwenol=E9.

Re: [B2-devel] Problem compiling JIT

[Gwenole B. <gb...@di...>]

Hi,

> What can I do about this? Should I download the sources from CVS, or
> do I need gcc 3?

Sorry, I only use gcc >=3D 3.2.2. You would have to fix and test =
yourself=20
support for gcc2.95 as I don't use so ancient compilers. I don't quite=20=

understand the first error messages either.

Bye,
Gwenol=E9.=

[B2-devel] Problem compiling JIT

[Jeffery T. <jef...@gm...>]

(I originally sent this when I wasn't subscribed to the list; I hope 
this doesn't cause duplicate copies.)

I'm trying to compile SheepShaver on a Pentium II, running Debian 3.0 
(Woody) with gcc 2.95.4, using Gwenole's snapshot
SheepShaver-2.2-20040609.tar.bz2. After coaxing configure to enable 
JIT (changing "#include <limits>" to "#include <limits.h>" in 
configure.in) and changing a command option to work with gcc 2.95
(-finline-limit=10000 to -finline-limit-10000), I get the following 
errors:

c++ -I../kpx_cpu/include -I../kpx_cpu/src -DUSE_JIT -I../include -I. -
DHAVE_CONFIG_H -D_REENTRANT -DDATADIR=\"/usr/local/share/SheepShaver\ 
" -g -O2  -I/usr/X11R6/include -I/usr/include/gtk-1.2 -I/usr/include/ 
glib-1.2 -I/usr/lib/glib/include -I/usr/X11R6/include -fomit-frame-po 
inter -mpreferred-stack-boundary=2 -malign-functions=0 -finline-limit-
10000 -c ../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp -o obj/basic-dyn 
gen-ops.o
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:40: global register 
variable follows a function definition
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:42: global register 
variable follows a function definition
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:44: global register 
variable follows a function definition
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:46: global register 
variable follows a function definition
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp: In function `void 
op_execute(uint8 *, basic_cpu *)':
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:300: jump to label 
`dummy_label1'
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:300:   from here
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:300:   crosses 
initialization of `const void * label2'
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp: In function `void 
op_jmp_slow()':
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:331: jump to label 
`dummy_label1'
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:331:   from here
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:331:   crosses 
initialization of `const void * label2'
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp: In function `void 
op_jmp_A0()':
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:345: jump to label 
`dummy_label1'
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:345:   from here
../kpx_cpu/src/cpu/jit/basic-dyngen-ops.cpp:345:   crosses 
initialization of `const void * label2'
make: *** [obj/basic-dyngen-ops.o] Error 1

What can I do about this? Should I download the sources from CVS, or 
do I need gcc 3?

Thanks,
Jeff

Re: [B2-devel] SDL target

[Gwenole B. <gb...@di...>]

Hi,

SDL graphics is now somewhat working for both Basilisk II and=20
SheepShaver. I still have to do some cleaning but at least differences=20=

are now minimal. I have built a MacOS X.2 SheepShaver binary and it is=20=

only 750 KB stripped with SDL libraries statically linked in. Problem:=20=

it's twice as slow as the Linux/X11 version. :-(

BTW, I will likely move display-independent code for NQD support to=20
gfxaccel.cpp in the top src dir. I think with very little changes, the=20=

BeOS version could use that too. Currently, NQD is duplicated in SDL/=20
and Unix/ directories, that's bad.

>> - cdrom: SDL?
>
> Only CD audio functions.

That's what I saw. :-( I will have to check how to minimally use OS X=20
libraries from Nigel's support code for MacOS X.

>> Have I forgot anything?
>
> Mouse/keyboard input: SDL ;-)

It turns out keyboard support for SDL is flaky (as with X11 keysyms)=20
unless you use Window Server dependent scancodes/keycodes or you are=20
using US only operating systems. I will add Quartz mappings but I only=20=

have a PBG4 so some keys will be missing.

> We might actually try to do InputSprockets support for devices like
> joysticks, too.

If we could at least do InputSprockets support for mouse, that would be=20=

nice. ;-) I tried to see several times ago where the problem was and I=20=

failed. e.g. in Unreal, the mouse pointer remains sticked to the upper=20=

left hand corner with no possible motion at all.

Bye,
Gwenol=E9.=

[B2-devel] SDL support started

[Gwenole B. <gb...@di...>]

Hi,

I have just committed enough SDL/graphics in windowed mode to the=20
Basilisk II tree. I can now build Basilisk II on MacOS X without an X=20
server within the Unix directory, configured with --enable-sdl-video.=20
That's useless as the proper MacOS X port from Nigel is bound to be=20
better integrated and faster.

I am new to SDL programming, so please read SDL/video_sdl.cpp for=20
caveats & TODOs. ;-)

BTW, I tried to write some glue code for SheepShaver but it crashes.=20
:-/ It would be better to use the same OO design as for B2 but I am=20
lacking platforms (BeOS) to regression test on. So, I will try to fix=20
the glue.

It would be better to use keycodes in SDL too, so the plan is to=20
enhance Unix/keycodes with other entries that can be matched through=20
aliases. e.g.

The XFree86 Project, Inc
alias SDL x11
alias SDL dga
9       53      # Esc
[...]

Quartz
alias SDL Quartz
[to be done]

Problem: I don't know if this solution is good enough.

Bye,
Gwenol=E9.=

[B2-devel] MacOS 9.0.4 support in CVS

[Gwenole B. <gb...@di...>]

Hi,

I have just committed small fixes to boot MacOS 9.0.4. Yeah! ;-)

Notes:
- You will have to make clean ; make completely after cvs update
- I have not tested the new code on x86 (little endian) platforms yet.

Christian, could you please update your homepage accordingly?

Bye,
Gwenol=E9, who spent the afternoon in thousand lines of logs.

Fwd: [B2-devel] Assert(cur_mose != -1) error on start

[Gwenole B. <gb...@di...>]

Hi,

> Also in regard to theblackwindow error, this also occurs in 800x600,=20=

> but I
> might have a clue as to its origin as it does not occur if you disable=20=

> the
> JIT compiler.

Try to use current CVS code. It should be less subject to compilers=20
variations in "branch" synthetic instructions.

Bye,
Gwenol=E9.

[B2-devel] SheepShaver snapshot 2004/06/09

[Gwenole B. <gb...@di...>]

Hi,

Since CVS stabilised, I placed convenience sources (CVS snapshots) +=20
precompiled binaries at:
<http://gwenole.beauchesne.free.fr/sheepshaver/files/>
Files: SheepShaver-2.2-20040609.tar.bz2, SheepShaver-2.2-12*.rpm

Changes since snapshot 2004/02/24:
- Minor improvements to the generic JIT engine. SheepShaver now=20
performs at around 1/8th of native speed.
- Make JIT generated code "reentrant" thus speeding up 68k audio=20
processing a lot.
- Add Native QuickDraw acceleration routines for BitBlt (srcCopy),=20
FillRect.
- Add TUN/TAP device support. Try to make ethernet support code=20
little-endian and 64-bit clean.
- Add run-time depth switching code from Basilisk II. It now supports 1=20=

bpp to host screen depths.
- Add "idlewait" option so that SheepShaver doesn't eat up all CPU when=20=

MacOS is not doing anything.
- Add experimental "jit68k" option to enable MacOS 68k DR emulator.
- Better fullscreen DGA support.
- Correctly setup XPRAM for first-time users.

Bye,
Gwenol=E9.=

Re: [B2-devel] SheepShaver and PRECISE_TIMING

[Christian B. <Chr...@un...>]

Hi!

On Fri, Jun 04, 2004 at 08:54:26PM +0200, Gwenole Beauchesne wrote:
> What PRECISE_TIMING stuff is supposed to do on BeOS/PPC?

Replace the 60Hz granularity with which TimerInterrupt() is normally called
(by INTFLAG_VIA) with a proper, high-resolution timer interrupt
(INTFLAG_TIMER).

> i.e. what are the visible effects/tests I could do in case I wanted the
> same for Unixish systems?

Enqueue time tasks with different durations and compare the actual elapsed
times obtained from Microseconds(). Without PRECISE_TIMING you should see
the 60Hz quantization. With PRECISE_TIMING it should be better.

Bye,
Christian

-- 
  / Physics is an algorithm
\/ http://www.uni-mainz.de/~bauec002/