[BusyBox] [PATCH] New comm applet

Larry Doolittle ldoolitt at recycle.lbl.gov
Wed Apr 6 17:12:05 UTC 2005


On Wed, Apr 06, 2005 at 05:56:30PM +0200, Rainer Weikusat wrote:
> Rob Landley <rob at landley.net> writes:
> >
> > I'm saying the demographic, if any, is a rounding error that may safely be 
> > ignored,
> 
> As I have already written: You are trying to push your
> 'Anti-Abstraction agenda' (cf STDIN_FILENO) by claiming that no
> non-ANSI terminals exist. This is an untrue statement and will remain
> untrue, [chop]

Calm down, guys.

This last comment nearly reaches the core issue:

1. Properly implemented, abstraction is good, it adds
   flexibility and lets you "say what you mean".
2. Poorly done abstraction is bad (turns into bloat).

It Would Be Nice if all busybox functions that manipulate
the screen (clear, reset, ls, vi, cmdedit, more) did so
through an abstraction layer.  But only if that abstraction
layer could be tuned for low bloat, a condition that (sadly)
curses, termcap, and terminfo fail.

I bet Rainer could win Rob over by writing a non-bloated
screen abstraction layer, and porting busybox to use it.
Unfortunately, that's a lot of work, and most people
(back to the demographic thing) would not benefit, since
the existing hard-coded ANSI strategy works fine for them.

I think you two should agree that the existing situation
(no abstraction) is less than ideal, but works well for
almost everybody.  Until and unless someone steps forward
with real, working, non-bloated abstraction code, we can
stop arguing and get back to work.

      - Larry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.busybox.net/pipermail/busybox/attachments/20050406/c2b09839/attachment-0002.pgp 


More information about the busybox mailing list