From: davidsen@ariel.crd.GE.COM (william E Davidsen)
Newsgroups: comp.os.linux
Subject: Good news, bad news
Date: 22 Jul 92 13:58:55 GMT
Reply-To: davidsen@crd.ge.com (bill davidsen)
Organization: GE Corporate R&D Center, Schenectady NY
Nntp-Posting-Host: ariel.crd.ge.com


  The good news is that I ported BPE (Binary Patch Editor) to Linux late
night and uploaded it to tsx-11 this morning. I included the binary,
source, and of course man page. Hopefully no more booting DOS to edit
binary files or floppy disks.

  The bad news is that future ports of software are going to come (from
me) slowly, since linux doesn't look as much like unix as I had
expected. Programs which compile without problem using gcc on SunOS and
V.4/386 generate thousands of warnings and errors on linux, due to the
changes in header files (or maybe because of 2.2.2 vs 2.1). Before
someone asks, yes, roughly 1.7 lines of error per line of code!

  I'll release things as I get the chance, but after four hours of
wading through error messages from one program which compiles on SunOS,
Ultrix, V.4, xenix, V.3, UNICOS, and even SysIII (it's an old program),
I conclude that when you start to diddle with ioctl and IPC linux is
not very similar to known unix versions. I had hoped to contribute some
of the missing pieces, but it seems less likely now.
-- 
bill davidsen, GE Corp. R&D Center; Box 8; Schenectady NY 12345
	It never ceases to amaze me that otherwise rational people, able to
	understand calculus, compound interest, and the income tax form, can
	continue to believe that poker is a game of chance.

From: hlu@phys1.physics.wsu.edu (Hongjiu Lu)
Newsgroups: comp.os.linux
Subject: Re: Good news, bad news
Date: 22 Jul 92 20:53:11 GMT
Organization: Washington State University

In article <1992Jul22.135855.17293@crd.ge.com>, davidsen@ariel.crd.GE.COM 
(william E Davidsen) writes:
|> 
|>   The bad news is that future ports of software are going to come (from
|> me) slowly, since linux doesn't look as much like unix as I had
|> expected. Programs which compile without problem using gcc on SunOS and
|> V.4/386 generate thousands of warnings and errors on linux, due to the
|> changes in header files (or maybe because of 2.2.2 vs 2.1). Before
|> someone asks, yes, roughly 1.7 lines of error per line of code!

Try gcc -traditional when 2.2.2d is release to public or contact me. You can
also try to convert it to ANSI C if you really want to use gcc 2.2.2. That
was what I did before I fed up with it and made some changes in 2.2.2d.

|> 
|>   I'll release things as I get the chance, but after four hours of
|> wading through error messages from one program which compiles on SunOS,
|> Ultrix, V.4, xenix, V.3, UNICOS, and even SysIII (it's an old program),
|> I conclude that when you start to diddle with ioctl and IPC linux is
|> not very similar to known unix versions. I had hoped to contribute some
|> of the missing pieces, but it seems less likely now.

For tty, Linux is POSIX. For IPC, you can only use Unix domain socket for now.
If you don't use those SYSV IPC, it should be very easy to port it to Linux. Do
you think that code is much more complicated than X11R5?


-- 
H.J.
Gcc/libc maintainer for Linux.

From: davidsen@ariel.crd.GE.COM (william E Davidsen)
Newsgroups: comp.os.linux
Subject: Re: Good news, bad news
Date: 23 Jul 92 16:10:59 GMT
Reply-To: davidsen@crd.ge.com (bill davidsen)
Organization: GE Corporate R&D Center, Schenectady NY
Nntp-Posting-Host: ariel.crd.ge.com

In article <1992Jul22.205311.7687@serval.net.wsu.edu>, hlu@phys1.physics.wsu.edu 
(Hongjiu Lu) writes:

| Try gcc -traditional when 2.2.2d is release to public or contact me. You can
| also try to convert it to ANSI C if you really want to use gcc 2.2.2. That
| was what I did before I fed up with it and made some changes in 2.2.2d.

  I do appreciate the thought (and thanks to those who suggested other
header files by mail), but this program really is pretty portable, and
compiles on the AT&T V.4 compiler, gcc 2.1 -ansi -pedantic, and SCO -W3
without problems.

| For tty, Linux is POSIX. For IPC, you can only use Unix domain socket for now.
| If you don't use those SYSV IPC, it should be very easy to port it to Linux. Do
| you think that code is much more complicated than X11R5?

  Thanks, this clarifies some of the problems, I will shelve this port
until linux has a bit more IPC (and might even look at generating a few
patches for some of the stuff I need). As for your question, I want to
use shared memory, pipes, SysV semiphores and message queues. I think
it's fair to say that the code is reasonably demanding in the IPC area,
perhaps more than X11R5. That doesn't relate 1:1 with complexity, of
course.

  By adding enough ifdefs I think I can use the file and termio stuff
from various parts already written, although the big blocks of BSD vs
SysV vs streams will have to be broken up.

  Everyone has been most helpful.
-- 
bill davidsen, GE Corp. R&D Center; Box 8; Schenectady NY 12345
	It never ceases to amaze me that otherwise rational people, able to
	understand calculus, compound interest, and the income tax form, can
	continue to believe that poker is a game of chance.