[cfe-dev] [LLVMdev] RFC: Upcoming Build System Changes

[Daniel Dunbar]

On Fri, Oct 28, 2011 at 2:12 PM, Christopher Jefferson
<chris at bubblescope.net> wrote:
>
> On 28 Oct 2011, at 18:26, Daniel Dunbar wrote:
>
>> On Fri, Oct 28, 2011 at 9:56 AM, Óscar Fuentes <ofv at wanadoo.es> wrote:
>>>
>>> You guys are mixing several things on the discussion. What Dan proposed
>>> makes no difference for Xcode. If it speeds up the Makefile-based build
>>> (something I doubt) that's good, but it is unnecessary for that system
>>> to fiddle with the cmake build as long as cmake can deal with the
>>> changes on its own.
>>
>> My proposal makes it relatively easy for a motivated engineer to
>> generate "clean" Xcode projects for LLVM.
>
> Couldn't a similarly motivated engineer work on greatly improving the "clean" Xcode projects for cmake in general?
>
> My personal main experience of a similar system is 'boost::build', or bjam, the build system used for the C++ boost libraries.
>
> This has slowly grown in scope, and I find completely unusable. It makes simple using boost, or configuring it slightly non-standardly, almost impossible without asking on IRC or mailing lists. bjam is the sole reason I have not submitted significant code to boost.
>
> Do not underestimate the benefit of hundreds of webpages written about CMake and make, detailing their nastier features. Where will your system's nastier features be documented?
>
> Sorry to sound negative, but I really don't want to have to learn yet another build system.

You don't. Nothing about what I am proposing changes how you build LLVM.

> One point which I didn't find answered, for casual developers.

See my original post, at the bottom.

>
> I currently do a bit of bug-finding on clang and libc++. All I am interested in is statements around the complexity of:
>
> CXX=g++4-.6 ./configure --prefix=/llvm --enable-expensive-checks; make -j8; make install
>
> Will that change at all?

No.

 - Daniel

>
> Chris