I would like to see command completion go away TOTALLY.
is anyone else with me here?
but starting in scripts is a step.
-- adm
On 01/30/03, at 15:18, Void had this to say..
// I agree with taking it out.
//
//
// void
//
// On Thu, 30 Jan 2003, Chip Norkus wrote:
//
// >
// > Hi folks. I've been quiet for a while, life and all that. However, I've
// > bumped into something that has been a bit of a sore spot for a while.
// > Automatic command completion in scripts. I'd, personally, like to see it
// > banished from the face of the earth.
// >
// > If you're not familiar with the feature, it's the feature that turns "/ali"
// > into "/alias" (as long as no other aliases/commands start with 'ali').
// >
// > Before anyone gets too excited, I don't mean for command completion to go
// > away *on the command line*. That's a useful feature, and one which I think
// > should stay. What I'm talking about relates purely to the execution of
// > aliases in scripts.
// >
// > The problem, as I see it, is twofold:
// > #1: I think there are substantial speed benefits to be gained by removing
// > this code. It will obviate the need for sorting, and will make the use of
// > hash-based lookups trivial.
// > #2: I think epic is violating POLA (principle of least astonishment) by
// > doing this at all. Imagine the case of a program who does the following:
// >
// > alias ali {
// > echo auto litigation system enabled
// > # ...
// > }
// >
// > ali ...
// >
// > Now imagine if they accidently mis-type 'ali' (say, as 'alli') and execute
// > their script. One would *expect* that doing "ali ..." will result in a
// > "command not found" error. That's not the result, however. Assuming an
// > otherwise clean running area, the user's script will actually execute the
// > alias command with the given arguments.
// >
// > So what I'd like to do here is find out what people think of this proposal.
// > Of especial interest is the response from scripters about whether this is
// > good, bad, or neither, and also whether their scripts need any work for
// > this to be done. Basically, does anyone actually want this behavior to
// > remain, and if you do, can you explain why?
// >
// > If nobody actually wants to keep this, I would propose that as part of my
// > namespace work it be removed and, either then or at a later date, a new
// > lookup system for aliases/assigns be introduced.
// >
// > -wd
// >
//
// _______________________________________________
// List mailing list
// [EMAIL PROTECTED]
// http://epicsol.org/mailman/listinfo/list
//
_______________________________________________
List mailing list
[EMAIL PROTECTED]
http://epicsol.org/mailman/listinfo/list
