Boost logo

Boost :

Subject: Re: [boost] Alternative implementation for BOOST_PP_VARIADIC_SIZE
From: Paul Mensonides (pmenso57_at_[hidden])
Date: 2011-11-16 23:12:53

On Wed, 16 Nov 2011 04:18:00 -0500, Lorenzo Caminiti wrote:

> On Wed, Nov 16, 2011 at 3:13 AM, Jens Gustedt <jens.gustedt_at_[hidden]>
> wrote:
>> Jens Gustedt <jens.gustedt <at>> writes:

>>> Please, be more concrete and show me such an input.
>> since there is no reply so far to my questions I conclude that you
>> don't have any concrete example, what a pitty

I have not had time to look up whatever it is that you are doing. What I
know is that there are a finite number of ways which can be used to
interact with an argument and none can be safely (i.e. without errors)
used on all input (discounting pathological unbalanced parentheses) with
the sole exceptions of stringizing (which you can always do) and token-
pasting to a placemarker (which you can always do) neither of which yield
results which are detectable for this purpose.

Without having to looked at precisely what you think works, I suspect that
you are doing something that should produce errors with some types of
input but doesn't because the particular preprocessor(s) that you've used
are permissive instead of strict.

If you show what you are doing here, I can take it apart to show where it
doesn't work. Pardon if you have already shown that and I missed it.

> It is not uncommon for people on this ML to reply within 1-2 days or
> over ~1 week (that is because most people on this ML are also busy with
> work-- a project release, etc). Occasionally, a thread gets a back and
> forth of replies within hours but that is the norm only for library
> review threads.

Quite right.

Paul Mensonides

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