Boost logo

Boost :

From: Karl Nelson (kenelson_at_[hidden])
Date: 2000-08-22 13:49:37

> on 8/21/00 3:41 PM, Karl Nelson at kenelson_at_[hidden] wrote:
> My opinions on the questions Karl asked:
> > stream/manipulator like versus function like
> I prefer the manipulator style.
> I love Dave Abrahams's proposed way of handling the string vs. stream
> contexts. If we don't go that way, string_format is the name I'd chose. I
> really want to see Karl's "return" of Dave's last API "volley". I think we
> are close to something really nice here.

Dave Abrahams's suggestion does look workable. I was a bit
worried about there being an ambiguity with

  cout << format(s) << 1 << (format(s2)) << 2;

But it appears that placing the parenthesis caused an
error which removed the ambiguity. :-)

The format of strings by themselves would be something like

  string s= (format("%d != %d) << 1 << 2).str();

(note, no endf here as that would hide the .str() function.)

> > full charT versus just a ostream and wostream version
> I thing doing the full charT thing is a good idea, as long as it doesn't
> somehow make the whole implementation too tricky and sink the whole thing.
> This makes it much more likely to get into a future C++ library standard,
> for one thing.

charT will definitely get done at some point. Though resolving the
first set of problems is of the most immediate concerns.

I think I have enough ideas to start implementing something which
matches closely which what boost wants. (though I am certainly
open to more feedback.)


Boost list run by bdawes at, gregod at, cpdaniel at, john at