Daryle Walker <darylew@hotmail.com> writes:
On 8/22/05 9:12 AM, "David Abrahams" <dave@boost-consulting.com> wrote:
Daryle Walker <darylew@hotmail.com> writes:
On 8/19/05 10:31 AM, "David Abrahams" <dave@boost-consulting.com> wrote:
Daryle Walker <darylew@hotmail.com> writes:
[SNIP]
Besides the caveats, you can just dump all the mandatory source code into your IDE's project file (or make file) and don't worry Boost-Build.
It isn't true. Some libraries need specific compiler/linker options on specific platforms. [SNIP]
Maybe we should consider that (and the separate DLL thing) to be a bug?
What's the "separate DLL thing" anyway?
So you're giving me a choice between admitting there's a bug in my build specification or having my library fail to build on some platform?
I didn't know about the specific-options requirement. I feel we should strive for our code to work without tweaks (but still be compatible with tweaks that the user may add). I guess there's no way around it.
What build specification are you referring to?
Boost.Python's, for example.
Whatever you think of Boost.Build, I don't see how underkill can work, and I don't see how libs/xxx/src/... is much worse than src/xxx/...
If the source files are in one location, it's easier to find out if a *.hpp file has corresponding source. The user otherwise has to dig through a lot of directories and/or documentation.
Which source file corresponds to a header is not something that should usually be of concern to our users, and even when it is, it isn't worth reorganizing the Boost project structure for.
It is a concern if the user adds Boost sources to his/her project file directly (instead of linking to a library pre-built by our scripts).
No. Being able to find the sources that belong to a library is a concern, but then that's easy: libs/library-name/src/*
I think if we _require_ Bjam, Boost.Build, or similar to use Boost, then we have failed.
We don't require it, whether or not the sources are reorganized. -- Dave Abrahams Boost Consulting www.boost-consulting.com