Re: [Boost-build] [v2] Ticket #83

From: Johan Nilsson (r.johan.nilsson_at_[hidden])
Date: 2006-05-17 15:56:48

Ryan, Volodya,

basically I concur with what Ryan is saying (some brief comments below).

"Ryan Gallagher" <ryan.gallagher_at_[hidden]> skrev i meddelandet 
news:loom.20060517T201935-451_at_post.gmane.org...
> Vladimir Prus <ghost <at> cs.msu.su> writes:
>>
>> First of all, why do you have identically named files in subdirectories? 
>> For
>> now, I don't understand why this is a such a recurring problem for 
>> users -- 
>> we had several reports already.
>
> Hi Volodya,
>
> Your question had been asked on a separate thread, and I get the
> feeling that it's a hint that this is bad practice, but I have yet to hear
> a decent argument for that.  Let me explain my personal reasons, which I
> believe match with some other users.
>
> C++ allows you to separate your code by namespaces.  Like Boost code,
> our directory layout often matches that of the namespace.  Further
> our filenames often match that of our class names.

This is exactly what I'm doing most of the time. It might not be (well, it 
isn't) the way most Boost libraries are structured, but that doesn't 
necessarily mean it's bad. To get wider acceptance of Boost.Build as a 
general build tool, it should support other layouts as well.

>
> I believe that this is all considered good practice in general.  The
> questionable practice may be in having the same class name in multiple
> namespaces within the same project.

Wasn't this the very reason why namespaces were invented? I.e.; not having 
to use warts on names simply to disambiguate them.

[snip lots of good points from Ryan]

> My concern is simply that currently bbv2 does not seem to have a matching
> concept for namespaces.  I would love to discuss the design question on a
> different list, but as far as the topic of this list goes, I believe that
> this concern should be all that matters.
>
>> Implementation-wise, here are my thoughts:
>>
>> 1. For a Jamfile with a single source location, we can, for all source 
>> files
>> below that source location, retain the relative path, no problem.
>
> This is what I am requesting.

Ditto (I think).

[snip remaining stuff]

Thanks // Johan