There are two open issues with printf:

0.  Handling of 0-fill (eg %2.2d) is not standard (%02d).  (Oleg found a 
POSIX standard and it is different from what I was going by)

1.  It's slow, slow enough to matter sometimes.  I think the main source 
of slowness is in splitting the format string into fields, and I think 
that rewriting that part using dyad ;: would fix it.

These are low enough on my priority list that I encourage others to jump 
in and make the changes.

Henry Rich

On 2/20/2011 1:45 PM, Ian Clark wrote:
> Since learning recently about printf in J (...an old friend from C
> days) I find it solves a lot of problems I've developed ad-hoc
> solutions to.
>
> Result: I've abandoned my ad-hoc solutions.
>
> If it's not much used (by the system) maybe it ought to be?
>
> That's just by-the-by. I'm using it heavily in some middleware I'm
> developing. If it takes off it'll dragalong printf (...into the
> system?)
>
>
> On Sun, Feb 20, 2011 at 6:10 AM, Ric Sherlock<[email protected]>  wrote:
>> In J602 (and before) the printf.ijs script was available as part of
>> the system library
>>    fexist jpath '~system/main/printf.ijs'
>> 1
>>
>> So far it is not available in J701 (as part of the base system library
>> or as an addon).
>>
>> My preference would for it to be included as part of the base system
>> library as ~system/util/printf.ijs (like pacman.ijs) that way other
>> utils can depend on the fact that it is available to the user.
>>
>> Do others agree? Any other points of view?
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
>>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to