oslib-devel — discussions about oslib development/bugs/ideas, and so on...

Re: [Oslib-devel] trying to compile

[Paolo G. <pao...@ti...>]

On Thu, 2003-07-10 at 00:22, Paul Dufresne wrote:
> G:\oslib\oslib\xlib>make install
> gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
> -D__VIRCSW
> __ -c xinfo.c
> In file included from xinfo.c:0:
> c:/djgpp/lib/djgpp.ver:1: sys/version.h: No such file or directory (ENOENT)
> make.exe: *** [xinfo.o] Error 1

that is quite strange, because there is no reference at all to
sys/version.h in all the Shark source code. usually the whole oslib does
not rely at all on the standard include files provided by djgpp!

bye

PJ
-- 
Paolo Gai <pao...@ti...>
Scuola S. Anna

[Oslib-devel] cygwin with linux.mk as config.mk

[Paul D. <du...@ho...>]

Slowly beginning to understand...
I have copied linux.mk to ../config.mk to use cygwin.

BTW here the versions I used:
GCC 3.2
as(binutils) 2.13.90
make 3.79.1

Now, make clean works fine.
But when it comes to doing make install, and it comes to the
examples directory:
make -C examples all
make[1]: Entering directory `/usr/oslib/oslib/examples'
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__LINUX__ -I.. -c 
mbdem
o.c
In file included from mbdemo.c:27:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
ld -Bstatic -Ttext 0x220000 -s -nostartfiles -nostdlib -L../lib/ ../lib/x0.o 
mbd
emo.o --start-group -lhc -lhm -lhx -lkl -lcons --end-group -o mbdemo.xtn
ld: warning: cannot find entry symbol _mainCRTStartup; defaulting to 
00220000
../lib/x0.o(.text+0x4e):x0.s: undefined reference to `halt'
../lib/x0.o(.text+0x95):x0.s: undefined reference to `stkbase'
../lib/x0.o(.text+0x9f):x0.s: undefined reference to `stktop'
../lib/x0.o(.text+0xc9):x0.s: undefined reference to `startup'
mbdemo.o(.text+0x228):mbdemo.c: undefined reference to `_alloca'
../lib//libhc.a(stdlib.o.b)(.text+0x22):stdlib.c: undefined reference to 
`__exit
'
../lib//libhx.a(x1.o.b)(.text+0x50):x1.c: undefined reference to `mbi'
../lib//libhx.a(x1.o.b)(.text+0x168):x1.c: undefined reference to `mbi'
../lib//libkl.a(init.o)(.text+0x2c1):init.c: undefined reference to 
`ll_abort'
../lib//libkl.a(event.o)(.text+0x1fe):event.c: undefined reference to 
`ll_timer'

../lib//libkl.a(event.o)(.text+0x3cb):event.c: undefined reference to 
`last_hand
ler'
../lib//libkl.a(time.o)(.text+0x3ea):time.c: undefined reference to 
`ll_abort'
../lib//libhx.a(xinit.o)(.text+0xc):xinit.c: undefined reference to 
`ll_irq_tabl
e'
../lib//libhx.a(xinit.o)(.text+0x3a):xinit.c: undefined reference to 
`ll_irq_tab
le'
../lib//libhx.a(xinit.o)(.text+0x59):xinit.c: undefined reference to 
`X86_get_FP
U'
../lib//libhx.a(xinit.o)(.text+0x72):xinit.c: undefined reference to 
`X86_fpu'
../lib//libhx.a(xinit.o)(.text+0xde):xinit.c: undefined reference to 
`init_TR'
../lib//libhx.a(xsystab.o)(.text+0xb):xsystab.c: undefined reference to 
`IDT'
../lib//libhx.a(xsystab.o)(.text+0x19):xsystab.c: undefined reference to 
`IDT'
../lib//libhx.a(xsystab.o)(.text+0x21):xsystab.c: undefined reference to 
`IDT'
../lib//libhx.a(xsystab.o)(.text+0x7d):xsystab.c: undefined reference to 
`GDT_ba
se'
../lib//libhx.a(xsystab.o)(.text+0xb1):xsystab.c: undefined reference to 
`GDT_ba
se'
../lib//libhx.a(idtinit.o)(.text+0x11):idtinit.c: undefined reference to 
`h0'
../lib//libhx.a(idtinit.o)(.text+0x25):idtinit.c: undefined reference to 
`h1'
../lib//libhx.a(idtinit.o)(.text+0x39):idtinit.c: undefined reference to 
`h2'
../lib//libhx.a(idtinit.o)(.text+0x4d):idtinit.c: undefined reference to 
`h3'
../lib//libhx.a(idtinit.o)(.text+0x61):idtinit.c: undefined reference to 
`h4'
../lib//libhx.a(idtinit.o)(.text+0x75):idtinit.c: undefined reference to 
`h5'
../lib//libhx.a(idtinit.o)(.text+0x89):idtinit.c: undefined reference to 
`h6'
../lib//libhx.a(idtinit.o)(.text+0x9d):idtinit.c: undefined reference to 
`exc7'
../lib//libhx.a(idtinit.o)(.text+0xb1):idtinit.c: undefined reference to 
`h8'

...
...

../lib//libhx.a(idtinit.o)(.text+0x13e9):idtinit.c: undefined reference to 
`h254
'
../lib//libhx.a(idtinit.o)(.text+0x13fd):idtinit.c: undefined reference to 
`h255
'
../lib//libhx.a(ccpu.o)(.text+0x61):ccpu.c: undefined reference to 
`X86_is386'
../lib//libhx.a(ccpu.o)(.text+0x72):ccpu.c: undefined reference to 
`X86_hasCPUID
'
../lib//libhx.a(ccpu.o)(.text+0xb7):ccpu.c: undefined reference to 
`X86_isCyrix'

make[1]: *** [mbdemo.xtn] Error 1
rm mbdemo.o
make[1]: Leaving directory `/usr/oslib/oslib/examples'
make: *** [install] Error 2

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

[Oslib-devel] compiling ok on cygwin, but clean goes badly

[Paul D. <du...@ho...>]

I just had the idea of copying my oslib directory in cygwin, run a bash 
shell
and try make in the oslib directory...

It compiled without problems!

Now, I just have to study what I am supposed to do from there.

Trying make install after the make caused a problem:
...
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c abort.s
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c timeint.s
timeint.s: Assembler messages:
timeint.s:146: Warning: indirect ljmp without `*'
ar rs libkl.a mem.o cxsw-2.o init.o time.o aspace.o intevt.o event.o 
event1.o ab
ort.o timeint.o
make[1]: Leaving directory `/usr/oslib/oslib/kl'

Paul@paul-l8otc2nwe8 /usr/oslib/oslib
$ make install
make -C xlib install
make[1]: Entering directory `/usr/oslib/oslib/xlib'
md ..\lib
make[1]: *** [..\lib] Error 255
make[1]: Leaving directory `/usr/oslib/oslib/xlib'
make: *** [install] Error 2

Paul@paul-l8otc2nwe8 /usr/oslib/oslib
$ make clean don't work because there is no del command.

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

[Oslib-devel] DJGPP and Windows XP = ar problem?

[Paul D. <du...@ho...>]

I commented out the only line from <djgppdir>/lib/djgpp.ver which is now:
// include <sys/version.h>

...
This allowed me to go to:
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c vm86.c
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xbios.c
xbios.c: In function `x_bios_address':
xbios.c:53: warning: implicit declaration of function `GDT_read'
ar rs libhx.a xsys0.o cpu2.o exc.o ctx.o mem.o  xinfo.o x1.o xsystab.o 
xconv.o x
dosf.o xdosm.o ccpu.o fpu.o irq.o ctxsw.o xinit.o idtinit.o vm86.o xbios.o
c:\djgpp\bin/ar.exe: libhx.a: rename: Not enough memory (ENOMEM)
make.exe: *** [libhx.a] Error 1

Well, I begin to think the problem is that I am on Windows XP.
The computer have 512Mb so I doubt I really have not enough memory:

G:\oslib\oslib\xlib>go32-v2
go32/v2 version 2.0 built Dec 24 2001 21:25:02
Usage: go32 coff-image [args]
Rename this to go32.exe only if you need a go32 that can run v2 binaries as
well as v1 binaries (old makefiles).  Put ahead of the old go32 in your PATH
but do not delete your old go32 - leave it in the PATH after this one.
Set GO32_V2_DEBUG=y in the environment to get verbose output.

DPMI memory available: 238123 Kb
DPMI swap space available: 0 Kb

I had read something about problems of DJGPP with Windows XP.
Maybe that's what's happening with ar?


Now, I can't simply boot with FreeDOS and go on my hard-drive to compile, 
because
... because this is NTFS partition, I think...

So I am thinking... Maybe some space left on the disk to install oslib and 
DJGPP in a plain
FAT partition.

I like to see how problems comes in tree in programming world. :-)

Maybe just compiling myself binutils could help...

_________________________________________________________________
MSN Search, le moteur de recherche qui pense comme vous !  
http://fr.ca.search.msn.com/

[Oslib-devel] trying to compile

[Paul D. <du...@ho...>]

I downloaded your suggestion of files.
But then the same problem I had once upon a time:

G:\oslib\oslib\xlib>set djgpp
DJGPP=c:\djgpp\djgpp.env

G:\oslib\oslib\xlib>make install
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xinfo.c
In file included from xinfo.c:0:
c:/djgpp/lib/djgpp.ver:1: sys/version.h: No such file or directory (ENOENT)
make.exe: *** [xinfo.o] Error 1

G:\oslib\oslib\xlib>dir c:\djgpp\include\sys\version.h
Volume in drive C has no label.
Volume Serial Number is F483-204B

Directory of c:\djgpp\include\sys

03-06-99  13:22               367 version.h
               1 File(s)            367 bytes
               0 Dir(s)   1 064 767 488 bytes free

G:\oslib\oslib\xlib>

So, maybe there is something wrong with my installation but I don't see 
what.

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

Re: [Oslib-devel] trying to compile

[Paul D. <du...@ho...>]

>According to the manifest directory, I have BNU210, CSDPMI5B, DJDEV203,
>GCC2953B, MAK3791B, and some other packages that (I think) are not
>important for compiling oslib.
Thanks, I'll try that tomorrow and keep the list informed.

>Maybe the trick is just to set the DJGPP environment variable to the
>proper djgpp.env file?
I did it, the problem is with the "supposed file" libhx.a, which indeed is 
not in
the directory. Don't know if it is supposed to be part of the packages
you gived the version.

_________________________________________________________________
MSN Search, le moteur de recherche qui pense comme vous !  
http://fr.ca.search.msn.com/

Re: [Oslib-devel] trying to compile

[Luca A. <luc...@em...>]

Hi all,

I just resurrected my dos machine... I have DJGPP installed on it. It is
an old version, but it still works ;-) : the oslib cvs head compiles
flawlessy.

According to the manifest directory, I have BNU210, CSDPMI5B, DJDEV203,
GCC2953B, MAK3791B, and some other packages that (I think) are not
important for compiling oslib.

Maybe the trick is just to set the DJGPP environment variable to the
proper djgpp.env file?

			Luca
-- 
_____________________________________________________________________________
 Copy this in your signature, if you think it is important:
                                N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Migliaia di prodotti al prezzo che decidi tu. Fai subito la tua offerta!
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1650&d=9-7

Re: [Oslib-devel] Mailing list archive don't work

[Luca A. <luc...@em...>]

Hi Paul,

On Wed, 2003-07-09 at 08:27, Paul Dufresne wrote:
> I tried a few times since tomorrow to get to the mailing-list archive on the 
> web site,
> but it tell's that either the list name was misspelled or the list must have 
> just changed
> and to try in 2 or 4 hours.
I think this is a sourceforge problem: all the lists archives seem to be
periodically disabled.

Anyway, oslib-devel is a very low traffic mailing list, and I doubt that
you will find something useful in the archives...

			Luca
-- 
_____________________________________________________________________________
 Copy this in your signature, if you think it is important:
                                N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Metti il turbo alla tua casella di posta: 100 MB per inviare e ricevere tutto quello che vuoi. Clicca qui
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1624&d=9-7

Re: [Oslib-devel] trying to compile

[Luca A. <luc...@em...>]

Hi Paul,

On Wed, 2003-07-09 at 07:48, Paul Dufresne wrote:
> Well, first, I don't see your january 31 snapshot.
Go to
http://sourceforge.net/projects/oslib/
then "Files", and you should see it. Unfortunately, the web site is
still a little bit confused... I have to rewrite the "download" section.

> But I finally did a cvs checkout. (I had problems with empty file response, 
> but it seems temporary).
Yes, this is a well known problem with the sourceforge anonymous cvs...
Retrying generally solves the problem.

> I am trying to use DJGPP 2.04.
> 
> So I copied mk\gnu.mk to config.mk.
Ok...

> I did made some changes to get it go longer, I include all my log here 
> however,
> to give you an idea of what I did.
> 
> ===============  First try ==================
> G:\oslib\oslib>make
> make -C xlib all
> make[1]: Entering directory `G:/oslib/oslib/xlib'
> gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c xsys0.s
> gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c cpu2.s
> gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c exc.s
> g:/djgpp/tmp\ccejONXw.s: Assembler messages:
> g:/djgpp/tmp\ccejONXw.s:595: Warning: Indirect ljmp without `*'
> g:/djgpp/tmp\ccejONXw.s:693: Warning: Indirect ljmp without `*'
> g:/djgpp/tmp\ccejONXw.s:794: Warning: Indirect ljmp without `*'
Ok, these warnings are expected.
> gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c ctx.s
> g:/djgpp/tmp\ccd6QuXI.s: Assembler messages:
> g:/djgpp/tmp\ccd6QuXI.s:224: Warning: Indirect ljmp without `*'
> gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c mem.s
> gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
> -D__VIRCSW
> __ -c xinfo.c
> In file included from xinfo.c:0:
> g:/djgpp/lib/djgpp.ver:1: sys/version.h: No such file or directory (ENOENT)
> make[1]: *** [xinfo.o] Error 1
> make[1]: Leaving directory `G:/oslib/oslib/xlib'
> make: *** [all] Error 2
This looks like a problem in the gcc installation... Anyway, try to
remove -nostdinc from C_OPT, and let me know if this solves the problem.

[...]
> gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
> -D__VIRCSW
> __ -c vm86.c
> In file included from vm86.c:33:
> ../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
These warnings with gcc 3.2 are expected... Don't worry about them.
> gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
> -D__VIRCSW
> __ -c xbios.c
> In file included from xbios.c:27:
> ../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
> xbios.c: In function `x_bios_address':
> xbios.c:53: warning: implicit declaration of function `GDT_read'
> ar rs libhx.a xsys0.o cpu2.o  exc.o ctx.o mem.o  xinfo.o x1.o xsystab.o 
> xconv.o x
> dosf.o xdosm.o ccpu.o fpu.o irq.o ctxsw.o xinit.o idtinit.o vm86.o xbios.o
> process_begin: CreateProcess(g:\djgpp\bin\ar.exe, ar rs libhx.a xsys0.o 
> cpu2.o e
> xc.o ctx.o mem.o xinfo.o x1.o xsystab.o xconv.o xdosf.o xdosm.o ccpu.o fpu.o 
> irq
> .o ctxsw.o xinit.o idtinit.o vm86.o xbios.o, ...) failed.
This is definitly a DJGPP installation problem... Are you sure that you
correctly installed DJGPP? If I remember correctly, this includes
setting some environment variables.
I have a working DJGPP on my dos machine... I'll try it as soon as I
will have some spare time, but it used to work last time that I checked.
I do not remember the exact files that I installed, but I just followed
the instructions from
http://www.delorie.com

I know that Salvo was able to compile the oslib cvs head by using DJGPP;
I'll ask him his configuration and I'll post it here.

				Luca
-- 
_____________________________________________________________________________
 Copy this in your signature, if you think it is important:
                                N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
...Occhiali da sole? Da vista? Lenti a contatto?
Vieni a provare la convenienza su Occhialeria.it!
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=881&d=9-7

[Oslib-devel] Mailing list archive don't work

[Paul D. <du...@ho...>]

I tried a few times since tomorrow to get to the mailing-list archive on the 
web site,
but it tell's that either the list name was misspelled or the list must have 
just changed
and to try in 2 or 4 hours.

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

[Oslib-devel] trying to compile

[Paul D. <du...@ho...>]

Well, first, I don't see your january 31 snapshot.
But I finally did a cvs checkout. (I had problems with empty file response, 
but it
seems temporary).

I am trying to use DJGPP 2.04.

So I copied mk\gnu.mk to config.mk.

I did made some changes to get it go longer, I include all my log here 
however,
to give you an idea of what I did.

===============  First try ==================
G:\oslib\oslib>make
make -C xlib all
make[1]: Entering directory `G:/oslib/oslib/xlib'
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c xsys0.s
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c cpu2.s
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c exc.s
g:/djgpp/tmp\ccejONXw.s: Assembler messages:
g:/djgpp/tmp\ccejONXw.s:595: Warning: Indirect ljmp without `*'
g:/djgpp/tmp\ccejONXw.s:693: Warning: Indirect ljmp without `*'
g:/djgpp/tmp\ccejONXw.s:794: Warning: Indirect ljmp without `*'
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c ctx.s
g:/djgpp/tmp\ccd6QuXI.s: Assembler messages:
g:/djgpp/tmp\ccd6QuXI.s:224: Warning: Indirect ljmp without `*'
gcc -x assembler-with-cpp -D__GNU__ -I.. -D__VIRCSW__ -c mem.s
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xinfo.c
In file included from xinfo.c:0:
g:/djgpp/lib/djgpp.ver:1: sys/version.h: No such file or directory (ENOENT)
make[1]: *** [xinfo.o] Error 1
make[1]: Leaving directory `G:/oslib/oslib/xlib'
make: *** [all] Error 2
==============================================
Here, I overwrote bnu2951b.zip with bnu214b.zip.
================Second tried ======================
G:\oslib\oslib>make
make -C xlib all
make[1]: Entering directory `G:/oslib/oslib/xlib'
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xinfo.c
In file included from xinfo.c:0:
g:/djgpp/lib/djgpp.ver:1: sys/version.h: No such file or directory (ENOENT)
make[1]: *** [xinfo.o] Error 1
make[1]: Leaving directory `G:/oslib/oslib/xlib'
make: *** [all] Error 2

G:\oslib\oslib>
------------------------------------------------------------------------------
Here, I downloaded and overwrote gcc2953b with gcc322b.zip.
From there, I was able to go a bit longer:
================ Last try ==================
G:\oslib\oslib>make
make -C xlib all
make[1]: Entering directory `G:/oslib/oslib/xlib'
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xinfo.c
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c x1.c
In file included from x1.c:31:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xsystab.c
In file included from xsystab.c:25:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xconv.c
In file included from xconv.c:25:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xdosf.c
In file included from xdosf.c:24:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xdosm.c
In file included from xdosm.c:24:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c ccpu.c
In file included from ccpu.c:26:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c fpu.c
In file included from fpu.c:30:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c irq.c
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c ctxsw.c
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xinit.c
In file included from xinit.c:24:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c idtinit.c
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c vm86.c
In file included from vm86.c:33:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
gcc -Wall -O -finline-functions -fno-builtin -nostdinc -D__GNU__ -I.. 
-D__VIRCSW
__ -c xbios.c
In file included from xbios.c:27:
../ll/i386/mem.h:308:17: warning: multi-line string literals are deprecated
xbios.c: In function `x_bios_address':
xbios.c:53: warning: implicit declaration of function `GDT_read'
ar rs libhx.a xsys0.o cpu2.o exc.o ctx.o mem.o  xinfo.o x1.o xsystab.o 
xconv.o x
dosf.o xdosm.o ccpu.o fpu.o irq.o ctxsw.o xinit.o idtinit.o vm86.o xbios.o
process_begin: CreateProcess(g:\djgpp\bin\ar.exe, ar rs libhx.a xsys0.o 
cpu2.o e
xc.o ctx.o mem.o xinfo.o x1.o xsystab.o xconv.o xdosf.o xdosm.o ccpu.o fpu.o 
irq
.o ctxsw.o xinit.o idtinit.o vm86.o xbios.o, ...) failed.
make (e=87): The parameter is incorrect.
make[1]: *** [libhx.a] Error 87
make[1]: Leaving directory `G:/oslib/oslib/xlib'
make: *** [all] Error 2
-------------------------------------------------------------------------------------
G:\oslib\oslib>
G:\oslib\oslib>ar --version
GNU ar 2.14 20030612

Not sure what to think from here.
Maybe I should not use DJGPP alpha (2.04) ?

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

Re: [Oslib-devel] problems compiling on cygwin

[Luca A. <luc...@em...>]

Hi,

> WFM with linux/gcc 2.96, and with some old versions of DJGPP.
> gcc 3.x and DJGPP 3.x should work also; at least we get them working
> under hark, anyway oslib has some different patches so I really do not
> remember its state...
Last time I checked, it worked fine with gcc 2.9? (sorry, I do not
remember what ? is... ;-), and Juras Benes reported it to work with gcc
3.2 under linux.

			Luca
-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Hai bisogno di riorganizzare la tua agenda? Risolvi il problema con un nuovissimo palmare!
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=762&d=9-7

Re: [Oslib-devel] problems compiling on cygwin

[Luca A. <luc...@em...>]

Hi Paul,

> I am far from being very at ease with configuration files.
> So what I say, may be look easy stuff to others. In fact that's what I hope. 
> :-)
> 
> OSlib: 0.1
This one has some serious bugs... I suggest you to get the cvs head, or
at least the January 31st shapshot.

> 
> First going to xlib and typing make install, gives the error that
> ../config.mk did not exist, so I created one with just an empty line.
Uhmm... This can be the reason for all the errors ;-)
It is my fault: maybe writing the documentation in LaTeX was not that
good idea... Anyway, you have to pick up the correct configuration file
from the mk directory, and to copy it in config.mk.
Cygwin is still not supported (sorry, I have not a cygwin installation
for testing); if you use DJGGP, doing a copy
mk\gnu.mk config.mk
should be enough.

> Why am I compiling it in the first place?
> Because I want to prove myself that the bug I have with biosdisk in 
> freeDos32 comes from oslib.
Thanks for doing this: it will help a lot!!! You may want to try the
biosdemo and diskdemo examples, and to enable the debugging in xlib (by
adding a -D__LL_DEBUG__ in C_OPT).

			Thanks,
				Luca
-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Occhialeria.it
Scopri le migliori marche a prezzi imbattibili
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=879&d=9-7

Re: [Oslib-devel] problems compiling on cygwin

[Paolo G. <pao...@ti...>]

WFM with linux/gcc 2.96, and with some old versions of DJGPP.
gcc 3.x and DJGPP 3.x should work also; at least we get them working
under hark, anyway oslib has some different patches so I really do not
remember its state...

bye

PJ

On Wed, 2003-07-09 at 05:42, Paul Dufresne wrote:
> Well, I tried with DJGPP, I thought I'd use alpha.
> But I took an old gcc (2.95).
> But I took the bnu (binutils) from alpha, my mistake.
> It does the same as on cygwin, in my previous message.
> 
> as is even more recent, version 2.14, than the one I had on cygwin 2.13.90 
> If I remember correctly.
> 
> I'll try to download an old binutils, and see.
> 
> _________________________________________________________________
> MSN Messenger : discutez en direct avec vos amis !  
> http://messenger.fr.msn.ca/
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by: Parasoft
> Error proof Web apps, automate testing & more.
> Download & eval WebKing and get a free book.
> www.parasoft.com/bulletproofapps
> _______________________________________________
> Oslib-devel mailing list
> Osl...@li...
> https://lists.sourceforge.net/lists/listinfo/oslib-devel
-- 
Paolo Gai <pao...@ti...>
Scuola S. Anna

Re: [Oslib-devel] problems compiling on cygwin

[Paolo G. <pao...@ti...>]

there should be an "mk" directory inside the main oslib root tree.
Inside you'll find some configuration files.

try something like (under linux)

cd oslib
cp mk/linux.mk config.mk

enjoy,

PJ

On Wed, 2003-07-09 at 03:41, Paul Dufresne wrote:
> I am far from being very at ease with configuration files.
> So what I say, may be look easy stuff to others. In fact that's what I hope. 
> :-)
> 
> OSlib: 0.1
> 
> First going to xlib and typing make install, gives the error that
> ../config.mk did not exist, so I created one with just an empty line.
> 
> Then I got:
> $ make install
> as   -o xsys0.o xsys0.s
> xsys0.s: Assembler messages:
> xsys0.s:31: Error: invalid character '(' in mnemonic
> xsys0.s:35: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:36: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:39: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:40: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:41: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:42: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:43: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:44: Warning: rest of line ignored; first ignored character is `('
> xsys0.s:51: Error: invalid character '_' in mnemonic
> xsys0.s:63: Error: invalid character '_' in mnemonic
> xsys0.s:72: Error: invalid character '_' in mnemonic
> xsys0.s:80: Error: invalid character '_' in mnemonic
> xsys0.s:88: Error: invalid character '_' in mnemonic
> xsys0.s:96: Error: invalid character '_' in mnemonic
> xsys0.s:103: Error: invalid character '_' in mnemonic
> xsys0.s:111: Error: invalid character '_' in mnemonic
> make: *** [xsys0.o] Error 1
> 
> with:
> $ as --version
> GNU assembler 2.13.90 20030308
> Copyright 2002 Free Software Foundation, Inc.
> This program is free software; you may redistribute it under the terms of
> the GNU General Public License.  This program has absolutely no warranty.
> This assembler was configured for a target of `i686-pc-cygwin'.
> 
> I guess I should try DJGPP instead.
> 
> Why am I compiling it in the first place?
> Because I want to prove myself that the bug I have with biosdisk in 
> freeDos32 comes
> from oslib.
> 
> _________________________________________________________________
> MSN Messenger : discutez en direct avec vos amis !  
> http://messenger.fr.msn.ca/
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by: Parasoft
> Error proof Web apps, automate testing & more.
> Download & eval WebKing and get a free book.
> www.parasoft.com/bulletproofapps
> _______________________________________________
> Oslib-devel mailing list
> Osl...@li...
> https://lists.sourceforge.net/lists/listinfo/oslib-devel
-- 
Paolo Gai <pao...@ti...>
Scuola S. Anna

Re: [Oslib-devel] problems compiling on cygwin

[Paul D. <du...@ho...>]

Well, I tried with DJGPP, I thought I'd use alpha.
But I took an old gcc (2.95).
But I took the bnu (binutils) from alpha, my mistake.
It does the same as on cygwin, in my previous message.

as is even more recent, version 2.14, than the one I had on cygwin 2.13.90 
If I remember correctly.

I'll try to download an old binutils, and see.

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

[Oslib-devel] problems compiling on cygwin

[Paul D. <du...@ho...>]

I am far from being very at ease with configuration files.
So what I say, may be look easy stuff to others. In fact that's what I hope. 
:-)

OSlib: 0.1

First going to xlib and typing make install, gives the error that
../config.mk did not exist, so I created one with just an empty line.

Then I got:
$ make install
as   -o xsys0.o xsys0.s
xsys0.s: Assembler messages:
xsys0.s:31: Error: invalid character '(' in mnemonic
xsys0.s:35: Warning: rest of line ignored; first ignored character is `('
xsys0.s:36: Warning: rest of line ignored; first ignored character is `('
xsys0.s:39: Warning: rest of line ignored; first ignored character is `('
xsys0.s:40: Warning: rest of line ignored; first ignored character is `('
xsys0.s:41: Warning: rest of line ignored; first ignored character is `('
xsys0.s:42: Warning: rest of line ignored; first ignored character is `('
xsys0.s:43: Warning: rest of line ignored; first ignored character is `('
xsys0.s:44: Warning: rest of line ignored; first ignored character is `('
xsys0.s:51: Error: invalid character '_' in mnemonic
xsys0.s:63: Error: invalid character '_' in mnemonic
xsys0.s:72: Error: invalid character '_' in mnemonic
xsys0.s:80: Error: invalid character '_' in mnemonic
xsys0.s:88: Error: invalid character '_' in mnemonic
xsys0.s:96: Error: invalid character '_' in mnemonic
xsys0.s:103: Error: invalid character '_' in mnemonic
xsys0.s:111: Error: invalid character '_' in mnemonic
make: *** [xsys0.o] Error 1

with:
$ as --version
GNU assembler 2.13.90 20030308
Copyright 2002 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License.  This program has absolutely no warranty.
This assembler was configured for a target of `i686-pc-cygwin'.

I guess I should try DJGPP instead.

Why am I compiling it in the first place?
Because I want to prove myself that the bug I have with biosdisk in 
freeDos32 comes
from oslib.

_________________________________________________________________
MSN Messenger : discutez en direct avec vos amis !  
http://messenger.fr.msn.ca/

Re: [Oslib-devel] bugs etc.

[Luca A. <luc...@em...>]

Hi all,
 
>  I tried to compile BIOSDemo with DJGPP and it works ok,
>  while when I compile it with GCC 3.x in Linux, when the demo finishes
>  the message about SP1!=SP2 is displayed. So I think there's an error
>  in GCC 3.x which produces a buggy code.

once again: I don't think this is a bug. In my opinion, it is only an
optimization. It seems to me that gcc 3.* likes to play strange games
with the stack, and we cannot assume ESP to be the same at the beginning
and at the end of the program...

If you want to do the check for debugging purpose, you will have to
compile with -O0 (some modifications to the cvs head are needed to
compile with -O0, and I'll commit them in the next days)...

Please, someone correct me if I am wrong.

				Luca
-- 
_____________________________________________________________________________
 Copy this in your signature, if you think it is important:
                                N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Vai contro corrente! Inizia a mantenerti in forma da Natale: Oliviero Città dello Sport!
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1074&d=1-6

[Oslib-devel] bugs etc.

[Juras B. <yb...@tu...>]

Hello oslib-devel,

 I tried to compile BIOSDemo with DJGPP and it works ok,
 while when I compile it with GCC 3.x in Linux, when the demo finishes
 the message about SP1!=SP2 is displayed. So I think there's an error
 in GCC 3.x which produces a buggy code.

-- 
Best regards,
 Juras                          mailto:yb...@tu...

Re: [Oslib-devel] biosdemo, vm86 etc.

[Luca A. <luc...@em...>]

Hi all,

On Sat, 2003-05-31 at 03:43, Juras Benesz wrote:
> Hello oslib-devel,
> 
>   Only now I noticed that biosdemo.xtn displays a message that
>   SP1!=SP2 when the program finishes. So there's a bug in callBIOS
>   code.

As I already wrote more than one time, I think this is not a bug, but it
is due to some optimizations done by gcc 3.*. Can someone try to
reproduce the problem with gcc 2.9*?

In my experience, the stack check failure is not directly related to the
vm86 code (I obtained this "problem" also when vm86 is not used).

Of course, if someone else has a different theory or can provide a
patch...

			Luca
-- 
_____________________________________________________________________________
 Copy this in your signature, if you think it is important:
                                N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Silvia l'ha trovato! Ora tocca a te. Trova l'amore della tua vita su Meetic.it!
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1573&d=31-5

[Oslib-devel] biosdemo, vm86 etc.

[Juras B. <yb...@tu...>]

Hello oslib-devel,

  Only now I noticed that biosdemo.xtn displays a message that
  SP1!=SP2 when the program finishes. So there's a bug in callBIOS
  code.

-- 
Best regards,
 Juras                          mailto:yb...@tu...

[Oslib-devel] About the stack check in the oslib examples...

[Luca A. <luc...@em...>]

Hi all,

as someone of you may have noted, the stack check at the end of the
oslib programs from the "examples" directory sometime fails... In
particular, it was pointed out the failure in the recently posted
diskdemo.c.

After some investigation, I arrived to the conclusion that there is no
real failure: the fact is that in some situations (in particular, when
using gcc 3.* with -O), the compiler performs strange optimizations
changing the stack pointer... Hence, we cannot assume that the stack
pointer at the beginning and at the end of the program is the same (try
-O0 to remove such optimizations...  After changing some "extern inline"
in "static inline", the examples will compile, and will not complain
about the stack pointer anymore).

But I could be wrong... Hence, if someone wants to investigate the
problem he is more than welcome.

BTW, while compiling with -O0 I found some "static inline vs extern
inline" errors... I'll commit the fixes in the next days.


			Luca
-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
E' iniziata l'era del fax virtuale! Si chiama EmailFax IN e significa: affidabilità, semplicità, qualità e risparmio.
Clicca per saperne di più
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1570&d=22-5

[Oslib-devel] some answer about oslib

[Luca A. <luc...@em...>]

Hi all,

here are some ansers to questions that I recently received about oslib:
1) l1_irq_bind() can be used to associate a handler to a hardware
interrupt (mapped starting from PIC1_BASE and PIC2_BASE). l1_exc_bind()
can be used to assign handlers to exceptions. Both these functions use
l1_int_bind(), that associates a handler to a IDT entry. Look at
ll/i386/pic.h
2)In protected mode, interrupts from 0 to 31 are exceptions... It is not
possible to distinguish them from interrupts 0..31 reflected from vm86
mode... (I hope that reflecting interrupts 0...31 from vm86 mode is not
needed. Otherwise, we need to check the TSS of the task generating the
interrupt).
3) I see the stack check failure in the recently posted diskdemo... I am
investigating it, but I suspect it is a get_SP() failure caused by
recent compilers (in other words: I suspect the program is ok, but the
check fails...).

				Luca
-- 
_____________________________________________________________________________
 Copy this in your signature, if you think it is important:
                                N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Difendi la tua casella di posta dai virus e dallo spam, prendi Email.it Pro15 o Pro50 la casella professionale e sicura.
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1048&d=21-5

[Oslib-devel] Re: oslib general protection fault

[Luca A. <luc...@em...>]

Hi Juras,

(I cc the oslib-devel mailing list, because the e-mails that I write to
you keep bounching - there is some _serious problem_ with your mail
server... I hope you will read this e-mail in the oslib-devel archives)
 
>   I was confused by the biosdemo example when an 0x6D handler is set.
>   The topic is that we can map a real-mode INT instruction into a protected one
>   through GPF (General Protection Fault) handler.
No, oslib uses a different approach. The IOPL of the VM86 task is set so
that an INT instruction does not generate a GPF, but triggers the
corresponding interrupt (returning to ring 0) according to the PM IDT.
Hence, the interrupt handler only has to do some tricks on the VM86
stack to invoke the RM handler, and everything is done... No need for
monitors or emulators ;-)

> So I thought that in
>   OSLib a GPF handler is defined and that handler called automatically
>   a protected mode interrupt with the same number as needed in the
>   vm86task.
No, this is not the case... Because of the IOPL setting, no GPF is
generated... If you _really want_ a GPF to be generated, you must set
the IDT entry corresponding to your interrupt as invalid (in this way,
the INT instruction will generate a GPF, and you will have to handle
it).

>   I need to write a
>   simple vm86 monitor for BIOS to see where INT instruction is called and what
>   interrupts are needed, so I need a personal exception handler for
>   GPF which will decode INT instruction and act as it must be.
Ok, if you really want to do the emulation stuff, you have two
possibilities:
1) set all the IDT entries but the first 32 to 0 (so that they are
invalid, and a GPF is generated)
2) change the IOPL level in vm86_callBIOS()

However, I suggest to simply set the hanlders that you need as in
biosdemo.c... If an INT is invoked and you did not set the proper
handler, you will see something like "Unhandled INT or Exc <number>" on
the screen, and the program will stop. Then, you will know that you need
to handle that int ;-)
(try to comment out the l1_int_bind() call in biosdemo.c...)

			Luca
-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Miotti.it eyewear: qualità, assortimento e sopratutto spedizioni gratuite!
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1451&d=20-5

[Oslib-devel] Re: [FreeDOS-32-dev] int 0x15

[Luca A. <luc...@em...>]

Hi Juras,

unfortunately, I have not too much time to work on this, right now...
Can you try the attached program, modify it to reproduce your problem,
and send it back to the list, so that I can actually reproduce the bug?

That would help debugging a lot...

				Luca
-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !



--
Email.it, the professional e-mail, gratis per te: http://www.email.it/f

Sponsor:
Ricordatelo: gli altri vogliono truffarti sul web, dai dialer ai programmi
spia. Manuali.net ti consegna le armi per difenderti con oltre 700 trucchi:
Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=1187&d=14-5