Need shell scripters to test Bash patch

[William Park]

On Fri, Jan 09, 2004 at 10:54:38AM -0500, John Wildberger wrote:
> On January 8, 2004 02:29 pm, William Park wrote:
> 
> > But, I do use regular expression aweful lot, both in conditional test
> > and in splitting/extracting strings.  And, Bash source is better for
> > patching.  I mean, Zsh source is true spaghetti.  :-)
> Forgive my ignorance, but could you plese explain how nto go about to use your 
> patch to add to bash? I would be keen to play with it.
> What are the steps involved?
> John

To compile,
    wget ftp://.../bash-2.05b.tar.gz		--> Bash source tarball
    wget http://home.eol.ca/~parkw/bash.diff	--> my patch
    tar -xzf bash-2.05b.tar.gz
    mv bash-2.05b bash
    cd bash
    patch -p1 < ../bash.diff
    CFLAGS="-DPATTERN_MATCHING" ./configure 
    make
    strip bash
    ./bash
	--> Now you are in my Bash shell.
    exit
	--> back to your original shell.

To make it your login shell, 
    cp bash /usr/local/bin
    echo /usr/local/bin/bash >> /etc/shells
    chsh -s /usr/local/bin/bash loginname

Now, some of usage...

1.  help case
    help for
    help match
    help array
    help arrayfilter		--> it's Python thing.
    help arraymap		--> it's Python thing.
    help read
    help echo

2.  Regular expression operators =~ and !~ will work in all conditional
    tests, ie.
	[ string =~ regex ]
	[[ string =~ regex ]]
	test string =~ regex
    They are all implemented through my builtin command
	match string regex

    Eg.
	[ abc123 =~ '([a-z][0-9])' ] && array SUBMATCH
	[[ abc123 =~ '[a-z]+' ]] && array SUBMATCH
	[[ abc123 =~ '([0-9]+)' ]] && array SUBMATCH

3.  Regular expression case statement.  Shell normally expects glob
    pattern, ie.
	case "..." in
	    glob_pattern) ... ;;
	    ...
	esac

    So, to trigger regex, put matching single/double quotes as the first
    and the last character of the pattern.  Eg.
	    abc*)		--> glob
	    abc\*)		--> glob
	    abc*[0-9])		--> glob
	    'abc')		--> regex
	    ''abc'')		--> regex
	    'abc.*[0-9]')	--> regex

4.  Array is the big one.  The most useful features are splitting and
    extracting regex from string, ie. -e and -v option.  In fact, this
    is what got me started.

	a=()
	array -e '[a-z]+' a abc123xyz789  &&  array a		--> abc xyz
	array -e '[0-9]+' a abc123xyz789  &&  array a		--> abc xyz 123 789
	a=()
	array -v '_+' a abc__123_xyz__789  &&  array a		--> abc 123 xyz 789
	array -v '[a-z_]+' a abc__123_xyz__789  &&  array a	--> abc 123 xyz 789 123 789

    -j option would be useful for cutting like 'cut' but with string not
    just a character.  Eg.
	b="$a"
	array -j '..' b  &&  array b		--> abc..123..xyz..789..123..789
	b=() 
	array -j '..1' b $a  &&  array b	--> abc 23..xyz..789 23..789

    -f option is for fixed width fields, like Awk's FIXEDWIDTHS.  Eg.
	a=()
	array -f 3,4 a 1234567890  &&  array a		--> 123 4567

    -s and -r option are for sorting and reversing the array.  Eg.
	array -s a
	array -r a

-- 
William Park, Open Geometry Consulting, <opengeometry-FFYn/CNdgSA at public.gmane.org>
Linux solution for data management and processing. 
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml