Re: [boost] Review : format2, yet another 'printf class'

From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-06-29 12:12:52

At 11:42 AM 6/29/2001, Samuel Krempp wrote:

 >I updated the files, as format2 now reaches "version 1.0"
 >It is earned because it now has complete support of the desired printf
 >features ;-)
 >That is, I added the '+', and ' ' flags.
 >As well as an extra flag, '=' to indicate 'centered adjustment'.
 >
 >Also format now behaves correctly when a manipulator puts the stream in
 >internal adjustment (fill characters goes between sign/basefield and rest
 >of the number)
 >
 >So it is now possible to do
 >cout << format("[%=6s] [%+6s] [%+6s] [% 6s] [%+6s] [%s]\n") % 123
 >              * internal * setfill('W') % 234
 >              * internal * setfill('X') % -345
 >              * setfill('Y') % 456
 >              * setfill('Z') % -10
 >	      % "foo"
 >	      % "invisible extra argument" << "bar.\n";
 >which prints :
 >[  123 ] [+WW234] [-XX345] [YY 456] [ZZZ-10] [foo]
 >bar.

I missed the original discussions some months back, but just looking at 
your example above,

*  Very Interesting work!

*  The choice of arithmetic operators "*" and "%" is totally unsettling to 
my eye.  Let's try it with "<<":

cout << format("[%=6s] [%+6s] [%+6s] [% 6s] [%+6s] [%s]\n") << 123
               << internal << setfill('W') << 234
               << internal << setfill('X') << -345
               << setfill('Y') << 456
               << setfill('Z') << -10
               << "foo"
               << "bar.\n";

That seems ever so much better if you can pull it off.

*  Having a disappearing argument ("invisible extra argument") seems like a 
bad idea; most likely it is a coding error so should not be silent, IMO.

*  Like Dave (in another message), I really hate sticky formatting; it has 
caused me a lot of past grief.

I hope you will continue to develop "format".  Seems like a very nice 
addition indeed.

--Beman