|
Boost : |
From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-03-18 08:54:18
At 03:34 PM 3/16/2001, Mark Rodgers wrote:
>From: "Beman Dawes" <bdawes_at_[hidden]>
>>
>> >#define BOOST_VERSION 0102101
>> >
>> >and document that
>> >
>> > BOOST_VERSION % 100 is the sub-minor version
>> > BOOST_VERSION /100 % 1000 is the minor version
>> > BOOST_VERSION / 100000 is the major version
>> >
>> >checking whether 1.21.0 or later is then as easy as
>> >
>> >#if BOOST_VERSION >= 0102100
>>
>> I'll go with this approach, absent a killer argument against. But
minus
>> the octal constant:-)
>
>Any reason why couldn't also add
>
>#define BOOST_VERSION_MAJOR (BOOST_VERSION / 100000)
>
>Then there is still only one version number to increment on new
>releases, but we avoid the need for people to work out how to crack
>apart the version number if they really do require it.
Several of us were trying to discourage looking at just the minor version
number in #if's. There is some experience (with gcc IIRC) that separating
out the minor version gets abused. But I don't know if that experience
applies to us, since they didn't have the equivalent of BOOST_VERSION,
again IIRC.
--Beman
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk