For the Mac and two of my boxes   ksh -c 'print $((KSH_VERSION))'   returns 0.  
Those are all ksh93s+ with various dates, all still older then 2008-08-08.  One 
linux box returns back a nice date value (20090501).  But that is the one box 
I'm not supposed to be developing on.  I wonder when Apple will get around to 
updating ksh93 in OSX?  No time soon, I'd guess.  As for the linux boxes, it's 
time to buy my SA a beer (or two).



Using your example below all of my versions (Version M 1993-12-28 s+   ..to..  
JM 93t+ 2009-05-01 ) of ksh return

        :10=now:11=-20:11=-21:12=1:13=1:14=1:

Yet, in my test script the 93t returns 'aaa' as the optarg instead of '−20'.  I 
must have something not exactly right in my USAGE.  I'll keep working on it in 
my spare moments.  




Thanks for the insight, you've been a great help.
Eric



On 9 Feb 2010, at 15:03 , Glenn Fowler wrote:

> 
> this command (source below)
>       test-getopts --date=now --flag=aaa --flag=bbb --user --media --verbose
> prints
>       :10=now:11=-20:11=-21:12=1:13=1:14=1:
> for ksh93 versions as old as 2001-05-04
> modulo suboption bugs for these releases:
>       2007-11-15 :10=now:11=aaa:11=bbb:12=:13=:14=:
>       2007-11-20 :10=now:11=aaa:11=bbb:12=:13=:14=:
>       2007-11-27 :10=now:11=aaa:11=bbb:12=:13=:14=:
> 
> for releases since 2008-08-08 this
>       ksh -c 'print $((KSH_VERSION))'
> will print the ksh release version as YYYYMMDD so it can be used in arithmetic
> and lexicographic comparisons
> 
> cat > test-getopts <<'!'
> USAGE=$'
> [-]
>  [10:date?Set the delivery date to \adate\a.]:[date]
>  [11:flag?Set the foobar to \atype\a.]:[type]
>    {       
>      [20:aaa?Desc of this type.]
>      [21:bbb?And this does something else.]
>    }       
>  [12:user?Retrieve user data.]
>  [13:media?Retrieve media content.]
>  [14:verbose?Verbosity is good sometimes.]
> '           
> 
> while getopts "${USAGE}" OPT
> do    print -n -- ":$OPT=$OPTARG"
> done
> print ":"
> !
> 
<snip>

_______________________________________________
ast-users mailing list
[email protected]
https://mailman.research.att.com/mailman/listinfo/ast-users

Reply via email to