From: Albrecht Kleine (kleine@ak.sax.de)
Date: Mon Dec 11 2000 - 17:35:02 CET
Michele wrote:
> > > 2. run sed on /tmp/buffer
> > Okay, you get it.. It works!
>
> Great! You call pipes without libc, and do not use kernel headers
> in your compilation. So, I'm asking my self, how your code link
> the kernel via assembler? Did you jumped in some fixed location,
> or you use a x86 interrupt for the system-call?
Exactly, this is done by x86 interrupt 0x80 in Linux (similar
in BeOS 0x25 and somewhat changed in FreeBSD). Some more details
are written in Konstantin Boldyshev's really nice Linux Assembly HowTo.
> In a previous email of mine (just sent) I suggested to try with
> "ex", the VI engine. But if sed works, no problem. I thik ex and
> sed are equipollent, as far editing on the fly is concerned.
Yes, 'sed' is quite okay (*** see footnote in P.S.)
>
> So, we have now an E3 on the top of the tremendous power of
> the regular UNIX expressions? :-))
I hope so :)
OTOH I'm not a vi expert so you should test it very much yourself.
>
> Ehm: I think, a this stage, is not so difficult for You,
> to run anything from E3. So, why do not add the "|" function?
Here we're again at the 4096-byte-piping-problem.
Using sed we have a defined interface using a file as sed's input
but sed's output is captured by E3. To make a more general approach
we need again the unlimited bidirectional pipe I've mentioned
at Friday night running via both STDIN + STDOUT, but not limited
to some peanuts of 4 kbyte.
> Thank very much! My first Xmas present, this year ...
Will be a lot of work for you for program testing :-))
(I hope to be ready for releasing that e3-1.2 next weekend.)
Cheers,
Albrecht
PS:
Note *** In former East Germany before 1990 we have had a "SED" party,
that really was NOT okay - unlike /bin/sed ;-)
---------------------------------------------------------------------
To unsubscribe, e-mail: mulinux-unsubscribe@sunsite.auc.dk
For additional commands, e-mail: mulinux-help@sunsite.auc.dk
This archive was generated by hypermail 2.1.6 : Sat Feb 08 2003 - 15:27:17 CET