Boost logo

Boost :

From: Jaakko Jarvi (jajarvi_at_[hidden])
Date: 2001-12-17 17:10:27


> This is totally cool. However, I have a philosophical problem with the
> preprocessor library (maybe philosophical is the wrong word): If I have to
> debug code generated by the preprocessor library, or even understand the
> source, I'm going to have a hard time.

This is a real problem. As a kind of an experiment, I used PREPROCESSOR in
lambda for one file which has two dimensional repetition:

some_template_1, some_template_2, etc.

within these templates the repetition is within the argument lists.

The code is much shorter than typing out all the
different versions. And a sincle preprocessor constant can be used to
change how many templates to generate.

And the library is awesome!

But still. The resulting code is _really_ hard to understand. And if you
get a compile time error, I guess most compilers just point you to the macro
invocation, which doesn't help a lot if the macro hides 500 lines of
code. So I didn't went a head and use preprocessor throughout the whole
library, but rather I'm thinkin to move back to the repetitive code in
the test case as well.

Jaakko


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk