On Thu, Jun 18, 2009 at 09:20:58PM +0200, Sam Ravnborg wrote: > The linux kernel is one of the major customers of > sparse and several sparse hackers are very > active on the kernel front too. > > How about combining forces and move sparse.git inside > the kernel. > > Benefits doing so: > > - Easy access to sparse from within the kernel. > Doing a make C=2 run is now trivial for everyone > > - We (the kernel hackers) has the source readily > avalibale when we need to hack it. > > Drawbacks: > > - We need to come up with a solution so it is > easy to distribute sparse binaries/source packages > without the need to distribute the full kernel source > > > The idea has come up before, and now that the kernel > has a tools/ directory we have a place where sparse can live. > > And I think Linus can find a way to pull sparse.git > with full history so we do not loose anything doing so. I think that Sparse needs to exist as a separate project for the benefit of all the non-kernel users of Sparse, but at the same time I think having Sparse in the kernel tree would help kernel developers actually run Sparse as part of their build process. Fortunately, Git offers a good solution to this problem: submodules. linux-2.6 could have Sparse as a submodule under the tools/sparse directory. Anyone who checked out linux-2.6 via git could then near-automatically get a copy of Sparse via git, and release tarballs of Linux could fairly easily include Sparse. The way submodules work, a commit in the linux-2.6 repository would refer to a specific commit in the sparse repostory, so linux-2.6 could always point to a version of Sparse known to more-or-less work with that version of Linux. How does that sound? - Josh Triplett -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
