Re: [v2] bad dependency analysis

From: David Abrahams (dave_at_[hidden])
Date: 2005-06-27 08:30:48

Vladimir Prus <ghost_at_[hidden]> writes:

> On Sunday 19 June 2005 02:21, David Abrahams wrote:
>> David Abrahams <dave_at_[hidden]> writes:
>> > Rene Rivera <grafik.list_at_[hidden]> writes:
>> >> David Abrahams wrote:
>> >>> I have a case where changing a header file doesn't cause a rebuild of
>> >>> the .cpp that includes it. Is there something in particular I should
>> >>> try to help debug it?
>> >>
>> >> BBv1 or BBv2?
>> >
>> > Sorry, v2. And I'm not sure I can repro this one anymore
>>
>> Well, it just happened to me again in a different context. It's an
>> easy repro, too. Just go to libs/mpl/test and build the "set" target
>> with your favorite compiler. Then
>>
>> touch boost/boost/mpl/set/aux_/set0.hpp
>>
>> Then build the set target again.
>
> I suspect it's the same problem with includes via macro. Looking at 
> preprocessor output, it looks like line 36 of boost/mpl/set.hpp, that reads:
>
> # include BOOST_PP_STRINGIZE(boost/mpl/set/AUX778076_SET_HEADER)
>
> affect this case.

Well, I guess we could at least make BB's scanner smart about some
macros. That was, after all, one of the original FTJam features; We
could do it entirely in bjam code but I bet we even have the builtin
function from FTJam still.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com