Mail Archives: djgpp/2005/10/13/12:17:18
| X-Authentication-Warning: | delorie.com: mail set sender to djgpp-bounces using -f
|
| From: | Sterten AT aol DOT com
|
| Message-ID: | <66.61655516.307fdfa3@aol.com>
|
| Date: | Thu, 13 Oct 2005 12:04:51 EDT
|
| Subject: | Re: smallest set of files needed
|
| To: | djgpp AT delorie DOT com
|
| MIME-Version: | 1.0
|
| X-Mailer: | 9.0 SE for Windows sub 5007
|
| X-Spam-Flag: | NO
|
| Reply-To: | djgpp AT delorie DOT com
|
-------------------------------1129219491
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
>> But it's clear now, that GCC/DJGPP is not a good choice to do
>> this.
>Why? Too big?
too big, too many files, too complicated the installation,
too difficult to explain to someone who is not a programmer.
>> So, is there another compiler which someone can recommend for this
>> purpose ?
>
>Okay, I'll list the reduced files for OpenWATCOM. I'm not sure if
>it is really any smaller...
>
>> (easy to install, handles basic C-commands, few and short files,
>>compatible,
>> free from copyright/license etc.)
>
>I have yet to see a Public Domain C compiler. Everything has some
>type of restriction... And, if it doesn't, it is usually too
>incomplete to be useful.
C is the most standard language, I think. Used for most University
research. Lots of software has been written in C. So we should have
this compiler already. It should even be included in the OS.
If it isn't there, then something must be wrong with the
whole computer busyness. (IMO)
>> I would even accept if only a subset of C-commands is compiled
>> and speed is not so important.
>
>You can also use CIL (C Intermediate Language) (link below) to
>simplify C code as much as you want. CIL is a C to C "translator"
>used to eliminate coding errors but also reduces complexity and
>posix code to simpler code.
>
>>But it should have good compatibility
>> with GCC/DJGPP.
>
>Not likely, due to POSIX, unless you switch to Linux. There are
>alot of simple C compilers for Linux e.g., TCC by F. Bellard (links
>below).
I don't know about Linux. Maybe it was a mistake not to choose Linux
in the first place, but now I'm too lazy to change.
>Sincerely,
>
>Rod Pemberton
>
>
>To use the reduced file sets for WATCOM, you'll need to put these
>in a .bat file and run it:
>SET WATCOM=C:\WATCOM
>SET EDPATH=C:\WATCOM\EDDAT
>SET INCLUDE=C:\WATCOM\H;C:\WATCOM\H\NT
>
>A minimal set of files for DOS RM OpenWATCOM:
>C:\watcom\binw\wcc.exe
>C:\watcom\binw\wcl.exe
>C:\watcom\binw\wlink.exe
>C:\watcom\binw\wlink.lnk
>C:\watcom\binw\wlsystem.lnk
>C:\watcom\binw\wstub.exe
>C:\watcom\h\*.h
>C:\watcom\h\sys\*.h
>C:\watcom\lib286\dos\clibs.lib
>C:\watcom\lib286\dos\emu87.lib
>C:\watcom\lib286\dos\graph.lib
>C:\watcom\lib286\math87s.lib
>
>A minimal set of files for DOS PM OpenWATCOM:
>C:\watcom\binw\wcc386.exe
>C:\watcom\binw\wcl386.exe
>C:\watcom\binw\wlink.exe
>C:\watcom\binw\wlink.lnk
>C:\watcom\binw\wlsystem.lnk
>C:\watcom\binw\wstub.exe
>C:\watcom\h\*.h
>C:\watcom\h\sys\*.h
>C:\watcom\lib386\dos\clib3r.lib
>C:\watcom\lib386\dos\clib3s.lib
>C:\watcom\lib386\dos\emu387.lib
>C:\watcom\lib386\dos\graph.lib
>C:\watcom\lib386\math387r.lib
>C:\watcom\lib386\math387s.lib
>C:\watcom\lib386\math3r.lib
>C:\watcom\lib386\math3s.lib
why don't they put these into one file ?
>Simple C compilers for DOS:
>
>smallC86 and smc88dos from here:
>http://www.retroarchive.org/cpm/archive/unofficial/download/
link doesn't work
>smc386c will could easily be retargetd to tasm, etc...
>http://www.physics.rutgers.edu/~vitchev/smallc-i386.html
smc386c.c doesn't compile with gcc3.2
>CIL (C Intermediate Language)
>CIL (C Intermediate Language) _http://manju.cs.berkeley.edu/cil/_
(http://manju.cs.berkeley.edu/cil/)
no compiler. And tons of documentation to read
>There is some limited 16bit support for DJGPP
>
>DJGPP's gcc patches/djlink/nasm
link doesn't work
>http://www.delorie.com/djgpp/16bit/djlink/
I don't understand this. What is it ? And it seems that you need
a compiler anyway.
>http://www.delorie.com/djgpp/16bit/
"This is not for the average programmer. This stuff does not install
easily and doesn't work well yet! Don't use unless you are prepared
to fix any bugs you find yourself."
>Simple C compilers floating around the internet that should compile
>for Linux:
no Linux
>Small C by Ron Cain (large number of microprocessors)
>FBCC and TCC by Fabrice Bellard
>BCC by Bruce Evans (from ELKS)
>Linux's Dev86 (from ELKS)
>OXCC C Compiler by Norman Culver
>PCC by C Ware Corporation
>Pacific C by Hitech Soft
>Amsterdam Compiler Kit
>Micro C by Dave Dunfield
>OSD (includes C compiler) by Chris Giese
Wouldn't it be nice if programmers could attach their source and
their compiler to the executables and people could easily edit
the executable - even non-programmers (e.g. output commands or
other simple things).
And then the changed executable could be run to recompile itself ?
Imagine this newsgroup-dialogue :
Q:
>can I change the size of the monster in the upper right corner
> in level 7 of game xyz ?
A:
> load xyz.exe into your editor and replace the 5 in line 745 by an 8
> and then recompile by running "xyz -compile". xyz.exe will change itself
> with the new monster included. You can recover the old version,
> "xyz -old"
-Guenter.
-------------------------------1129219491
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; charset=3DUS-ASCII">
<META content=3D"MSHTML 6.00.2800.1141" name=3DGENERATOR></HEAD>
<BODY id=3Drole_body style=3D"FONT-SIZE: 10pt; COLOR: #000000; FONT-FAMILY:=20=
Arial"=20
bottomMargin=3D7 leftMargin=3D7 topMargin=3D7 rightMargin=3D7><FONT id=3Drol=
e_document=20
face=3DArial color=3D#000000 size=3D2><FONT id=3Drole_document face=3DArial=20=
color=3D#000000=20
size=3D2>
<DIV> >> But it's clear now, that GCC/DJGPP is not a good choice=20=
to=20
do<BR> >> this.<BR> >Why? Too big?</DIV>
<DIV> </DIV>
<DIV>too big, too many files, too complicated the installation,<BR>too diffi=
cult=20
to explain to someone who is not a programmer.</DIV>
<DIV> </DIV>
<DIV> >> So, is there another compiler which someone can recommen=
d=20
for this<BR> >> purpose ?<BR> ><BR> >Okay, I'll l=
ist=20
the reduced files for OpenWATCOM. I'm not sure if<BR> >it is=20
really any smaller...<BR> ><BR> >> (easy to install, hand=
les=20
basic C-commands, few and short=20
files,<BR> >>compatible,<BR> >> free from=20
copyright/license etc.)<BR> ><BR> >I have yet to see a Publi=
c=20
Domain C compiler. Everything has some<BR> >type of=20
restriction... And, if it doesn't, it is usually=20
too<BR> >incomplete to be useful.</DIV>
<DIV> </DIV>
<DIV>C is the most standard language, I think. Used for most=20
University<BR>research. Lots of software has been written in C. So we should=
=20
have<BR>this compiler already. It should even be included in the OS.<BR>If i=
t=20
isn't there, then something must be wrong with the<BR>whole computer busynes=
s.=20
(IMO)</DIV>
<DIV> </DIV>
<DIV> >> I would even accept if only a subset of C-commands is=20
compiled<BR> >> and speed is not so=20
important.<BR> ><BR> >You can also use CIL (C Intermediate=20
Language) (link below) to<BR> >simplify C code as much as you=20
want. CIL is a C to C "translator"<BR> >used to eliminate codi=
ng=20
errors but also reduces complexity and<BR> >posix code to simpler=20
code.<BR> ><BR> >>But it should have good=20
compatibility<BR> >> with GCC/DJGPP.<BR> ><BR> >N=
ot=20
likely, due to POSIX, unless you switch to Linux. There=20
are<BR> >alot of simple C compilers for Linux e.g., TCC by F. Bellar=
d=20
(links<BR> >below).</DIV>
<DIV> </DIV>
<DIV>I don't know about Linux. Maybe it was a mistake not to choose Linux<BR=
>in=20
the first place, but now I'm too lazy to change.</DIV>
<DIV> </DIV>
<DIV> >Sincerely,<BR> ><BR> >Rod=20
Pemberton<BR> ><BR> ><BR> >To use the reduced file s=
ets=20
for WATCOM, you'll need to put these<BR> >in a .bat file and run=20
it:<BR> >SET WATCOM=3DC:\WATCOM<BR> >SET=20
EDPATH=3DC:\WATCOM\EDDAT<BR> >SET=20
INCLUDE=3DC:\WATCOM\H;C:\WATCOM\H\NT<BR> ><BR> >A minimal se=
t of=20
files for DOS RM=20
OpenWATCOM:<BR> >C:\watcom\binw\wcc.exe<BR> >C:\watcom\binw\=
wcl.exe<BR> >C:\watcom\binw\wlink.exe<BR> >C:\watcom\binw\wl=
ink.lnk<BR> >C:\watcom\binw\wlsystem.lnk<BR> >C:\watcom\binw=
\wstub.exe<BR> >C:\watcom\h\*.h<BR> >C:\watcom\h\sys\*.h<BR>=
>C:\watcom\lib286\dos\clibs.lib<BR> >C:\watcom\lib286\dos\e=
mu87.lib<BR> >C:\watcom\lib286\dos\graph.lib<BR> >C:\watcom\=
lib286\math87s.lib<BR> ><BR> >A=20
minimal set of files for DOS PM=20
OpenWATCOM:<BR> >C:\watcom\binw\wcc386.exe<BR> >C:\watcom\bi=
nw\wcl386.exe<BR> >C:\watcom\binw\wlink.exe<BR> >C:\watcom\b=
inw\wlink.lnk<BR> >C:\watcom\binw\wlsystem.lnk<BR> >C:\watco=
m\binw\wstub.exe<BR> >C:\watcom\h\*.h<BR> >C:\watcom\h\sys\*=
.h<BR> >C:\watcom\lib386\dos\clib3r.lib<BR> >C:\watcom\lib38=
6\dos\clib3s.lib<BR> >C:\watcom\lib386\dos\emu387.lib<BR> >C=
:\watcom\lib386\dos\graph.lib<BR> >C:\watcom\lib386\math387r.lib<BR>=
>C:\watcom\lib386\math387s.lib<BR> >C:\watcom\lib386\math3r=
.lib<BR> >C:\watcom\lib386\math3s.lib</DIV>
<DIV> </DIV>
<DIV>why don't they put these into one file ?</DIV>
<DIV> </DIV>
<DIV> >Simple C compilers for DOS:<BR> ><BR> >smallC=
86=20
and smc88dos from=20
here:<BR> >http://www.retroarchive.org/cpm/archive/unofficial/downlo=
ad/</DIV>
<DIV> </DIV>
<DIV>link doesn't work</DIV>
<DIV> </DIV>
<DIV> >smc386c will could easily be retargetd to tasm,=20
etc...<BR> >http://www.physics.rutgers.edu/~vitchev/smallc-i386.html=
</DIV>
<DIV> </DIV>
<DIV>smc386c.c doesn't compile with gcc3.2</DIV>
<DIV> </DIV>
<DIV> >CIL (C Intermediate Language)<BR> >CIL (C Intermediat=
e=20
Language) <A=20
href=3D"http://manju.cs.berkeley.edu/cil/">http://manju.cs.berkeley.edu/cil/=
</A></DIV>
<DIV> </DIV>
<DIV>no compiler. And tons of documentation to read</DIV>
<DIV> </DIV>
<DIV> >There is some limited 16bit support for=20
DJGPP<BR> ><BR> >DJGPP's gcc patches/djlink/nasm</DIV>
<DIV> </DIV>
<DIV>link doesn't work</DIV>
<DIV> </DIV>
<DIV> >http://www.delorie.com/djgpp/16bit/djlink/</DIV>
<DIV> </DIV>
<DIV>I don't understand this. What is it ? And it seems that you need<BR>a=20
compiler anyway.</DIV>
<DIV> </DIV>
<DIV> >http://www.delorie.com/djgpp/16bit/</DIV>
<DIV> </DIV>
<DIV>"This is not for the average programmer. This stuff does not=20
install<BR> easily and doesn't work well yet! Don't use unless you are=
=20
prepared<BR> to fix any bugs you find yourself."</DIV>
<DIV> </DIV>
<DIV> >Simple C compilers floating around the internet that should=20
compile<BR> >for Linux:</DIV>
<DIV> </DIV>
<DIV>no Linux</DIV>
<DIV> </DIV>
<DIV> >Small C by Ron Cain (large number of=20
microprocessors)<BR> >FBCC and TCC by Fabrice Bellard<BR> >B=
CC=20
by Bruce Evans (from ELKS)<BR> >Linux's Dev86 (from=20
ELKS)<BR> >OXCC C Compiler by Norman Culver<BR> >PCC by C Wa=
re=20
Corporation<BR> >Pacific C by Hitech Soft<BR> >Amsterdam=20
Compiler Kit<BR> >Micro C by Dave Dunfield<BR> >OSD (include=
s C=20
compiler) by Chris Giese</DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV><BR>Wouldn't it be nice if programmers could attach their source=20
and<BR>their compiler to the executables and people could easily edit<BR>the=
=20
executable - even non-programmers (e.g. output commands or<BR>other simple=20
things).<BR>And then the changed executable could be run to recompile itself=
=20
?</DIV>
<DIV> </DIV>
<DIV><BR>Imagine this newsgroup-dialogue : <BR>Q:<BR> >can I change=20=
the=20
size of the monster in the upper right corner<BR> > in level 7 of ga=
me=20
xyz ?<BR>A:<BR> > load xyz.exe into your editor and replace the 5 in=
=20
line 745 by an 8 <BR> > and then recompile by running "xyz=20
-compile". xyz.exe will change itself<BR> > with the new monst=
er=20
included. You can recover the old version, <BR> > "xyz -old"</DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV>-Guenter.</DIV></FONT></FONT></BODY></HTML>
-------------------------------1129219491--
- Raw text -