🚀 go-pugleaf

In article <41crmm$o3u@newsbf02.news.aol.com>,
ffarance@aol.com (FFarance) wrote:

>The "printf" and "scanf" funcitons are extended via the "?" type modifier.
>For example:
>
>	printf("%?d",precof(x),x);
>
>will always work.  Think of "?" as a combination of "*" (takes the next
>argument) and "l" and "h" (the type modifiers).  

In article <41g85d$kc@newsbf02.news.aol.com>,
ffarance@aol.com (FFarance) wrote:
>> From: geisler@mcs.anl.gov (Jonathan Geisler)
>> 
>> It seems like a little overkill to have to always include a ?/precof()
>> combination for any precision number.  Why not define a new modifier
>> that will automatically do the same lookup as precof() without the need
>> for the programmer to explicitly state it?  So, for example, the 2 above
>> examples become (assuming the new modifier is 'P'):
>> 
>>     printf("%P", x);                   AND
>>     printf("%*P", field_width, x);
>> 
>> This looks much more reasonable and readable to me.
>
>How would this lookup work?  

I can think of something in between these two proposals that might be
a little better.  You could use something like

#define vpi(x)  precof(x),(x)   /* variable precision int */

    printf("%D", vpi(x));

The format %D would specify two values being passed: an int specifying
the precision, followed by the actual value.  This might be in
addition to %?d.  Formats %u, %x, %o, and %i could be similarly
extended.

Although trickier, this might also be used for *scanf:

#define vpip(x) precof(*(x)),(x)    /* variable precision int ptr */

    sscanf (s, "%D", vpip (&x));

Thad