NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/amiga/misc-part2/12171

Internet Message Format | 1996-08-05 | 8.4 KB

Path: blue.cse.ogi.edu!leneis
From: [e-mail] (Tony Leneis)
Newsgroups: comp.sys.amiga.misc
Subject: Re: AmigaOS 4.x features
Date: 12 Apr 1996 23:43:55 GMT
Organization: Oregon Graduate Institute (OGI), Portland, Oregon
Message-ID: <4kmprr$[e-mail]>
References: <4j7ein$[e-mail]> <4jkg5j$[e-mail]> <4jl6pr$[e-mail]> <4kkf60$[e-mail]>
NNTP-Posting-Host: blue.cse.ogi.edu

Charlie Moody <[e-mail]> wrote:
>On 31 Mar 1996 05:56:11 GMT, Tony Leneis wrote:
>: test#a       <- Matches 'test', 'testa', 'testaa', 'testaaa', etc.
>
>  test*a       <- Matches 'test', 'testa', 'testaa', 'testaaa', etc.

test*a also matches 'testba', 'testbadfilea', 'test1234a', etc. while
test#a only matches test followed by zero or more a's.  It's not the
same thing.

>: test#(ab)c   <- Matches 'testc', 'testabc', 'testababc', 'testabababc', etc.
>
>Ya got me on this one...but do you (does anyone) *really* use such
>screwball naming conventions?

Not often, but it can be handy from time to time, especially when you're
dealing with a slew of data files that follow some sort of a regular
naming scheme.  A more common one for me would be something like 'data#[0-9]',
which selects any file named data followed by zero or more digits.

>: On the other hand, there are no UNIX globbing patterns which cannot be
>: expressed as an Amiga pattern.  Hence, Amiga patterns are more powerful.
>
>It's afloor wax!  It's a breath mint!  It's BOTH!!!  As I said...

I wasn't trying to sell it - I was simply trying to point out that the Amiga
patterns are more powerful than UNIX globbing patterns.  I don't have your
original post anymore, but as I recall you were complaining that the Amiga
pattern matching was not as powerful as what is provided by UNIX shells, which
just is not true.

>Please understand:  it's not my intention to insult my own machine
>(currently a much-upgraded 2000, replacing my grey-bearded 1000, which is
>now sunning itself in the back 40);  my comments were directed at making
>the *CLI* more flexible/convenient.  File globbing or not, I have always
>found the standard Amiga pattern-matching to be less simple, less
>intuitive, and hence less useful when moving bunches of files around
>(which I do...a lot....).

Point taken.  If that had been all your original post had said, I probably
would have left it alone.  I just had a problem with the 'not as powerful
as' and 'poorly designed' bits of the post.  The Amiga pattern matching
syntax is different from UNIX's file globbing, but then again it's also
different from VMS's and MS-DOS's...  It is not less powerful.

>: Would you have preferred the standard UNIX regular expression
>: syntax?  Then you'd have to type '.*' instead of '#?', and you'd have to
>: remember to escape the '.' whenever you didn't want it to act as the Amiga
>: '?' metacharacter.  (Kind of a bummer if you're looking for '#?.c', or should
>: I say '.*\.c' as a UNIX regular expression.)
>
>???  I've never encountered such a thing as you seem to be describing,
>from System III thru the current (Vr4?).  Certainly I've never had to
>jump thru *these* hoops on a unix box.

Note that I was talking about regular expressions there.  I don't know of any
UNIX shell that provides regular expressions for file expansion purposes, but
it's what many of the utilities use: awk; egrep and grep; ed, ex, sed, and vi;
and csplit to name a few.  This means that you have to learn two types of
pattern matching for UNIX.  The nice thing about the Amiga's pattern matching
is that it's sufficiently general to be used anywhere that regular expressions
would be useful, and it doesn't suffer from UNIX regular expression's
unfortunate choice of metacharacters.  (Unfortunate when working with common
filename pattern matching, anyway.)

>Or, to quote some old fart:  "It's not that the bear sings well:  it's
>that the bear sings at all."  Or another:  "Use the right tool for the job."

Exactly.  In my experience, I've found myself wishing that UNIX shells provide
the same functionality as the Amiga's pattern matching.  Actually, when you
get down to it, I think the pattern matching provided by WB 3.0 and above
pretty much does what you want, except that you have to use a utility to turn
on the map-*-as-#? function.  All the other metacharacters work the same,
except that you have to use '~' instead of '^' to compliment sets.
Incidentally, another bonus of 3.1 over 2.04 is all the ADOS commands accept
multiple filenames.  2.04 used to drive me nuts when I wanted to delete 3 or
4 files that were not practical to generate patterns for.

>: Just out of curiosity, whose wheel should they have
>: used instead of inventing their own?  As far as I know, there was no RFC or
>: other standard for file wildcarding when AmigaDOS was developed.  Which UNIX
>: shell should they have followed?
>
>This is just a *bit* disingenuous, doncha think?  Do you believe standards
>grow on RFC's?  Taking what works and improving it is a time-honored way
>of making progress.

Not really.  You complained that the Amiga's pattern matching was different
from everybody else's.  I was just wondering whose scheme they should have
followed.  Heck, no two UNIX shells agree on filename expansion, other than
'*', '?', and perhaps the '[]' pair (some shells don't know what to do with
'[^A-Z]'.)

>Much more useful FOR WHAT?  I say it's not more useful, in my experience,
>for moving masses of files around from the command line, because in my
>experience it's *not* more useful.  It may be just the thing for
>complementing complex scripted functions, but I don't care, really.  If I
>wanted to do enough studying to master that art, I'd tackle perl, or Python.

And my experiences have been just the opposite.  Then again, at work I have
to deal with tens of thousands of files that follow regular naming
conventions (e.g. an embedded date or sequence number).  There have been
many situations where I've really wanted ADOS's pattern matching power.
Fortunately, I was able to port GNU find to the UNIX boxes I work on, so
I'm able to do regular expression matching; however, typing things like

find . -regex '\./eft[0-9]*\.dat' -maxdepth 1 -exec rm {} \;

under UNIX is not as convenient as typing

delete eft#[0-9].dat

on the Amiga.  To be fair, find is an incredibly powerful command that can
do all sorts of cool things; however, I have to use it or write some custom
script that uses sed or grep or something in order to get regexp capable
pattern matching.  The shells I have available (sh, ksh, and bash) can't
hack it.

>Granted, I was working on unix boxes long before I got my 1000, and I'm
>*used* to doing certain things w/ the cli;  for me, it seems that the
>unix command line was designed as a primary interface, and the Amiga cli
>does not improve on it, and doesn't even equal it, in ease of common
>use.  I've managed to avoid MooseDOS for most of my career, and I'm very
>pleased about that;  and I don't *want* to run Linux or whatever:  I
>*like* my machine!

I'm just the opposite - I didn't start using UNIX with any frequency until
1987, so I'd already learned the Amiga's pattern matching.  I agree that
the Amiga CLI could use a makeover; however, I don't see anything really
wrong with the pattern matching.

>*sigh*  This wasn't meant to be a fight;  and as you say, the
>functionality I asked for is substantially available to me now on my
>Amiga, so this really is a moot point.

Agreed.  Really, the only thing I was attempting to point out was that
the Amiga pattern matching is not less powerful than what UNIX shells
provide, and that there is nothing wrong with the design.  It *is*
different, and the learning curve is steeper (because there are more
meta characters to learn), but that's it.  It's certainly not underpowered
or a kludge.  Personally, I feel that the manuals could use a little work
on explaining how Amiga pattern matching works, but I don't think it should
be scrapped.  If you think it should be redone because you don't care for
the syntax and semantics, that's fine.  I only objected to the incorrect
statement that AmigaDOS pattern matching was less powerful than UNIX shell
pattern matching being presented as fact.

-Tony
--
=====///===Hi!==I'm=a=.signature=virus!==Copy=me=to=your=.signature!===///==
__  /// [e-mail]          From the desk of:   Hillary,  __  ///
\\\///  [e-mail]       -->Tony Leneis<--   Gennifer, \\\///Only
 \XX/   [e-mail]  HMC '91  Go Stags!  AND NOW US!\XX/Amiga