On Mon, 26 Mar 2018 07:54:38 -0500, Michael Catanzaro <[email protected]> wrote: > On Sat, Mar 24, 2018 at 6:03 PM, Cadu Bentzen <[email protected]> > wrote: > > I included calling this script at the end of the build-webkit (under > > a command line option) script. So far it has work for me, but I > > wonder if this is the right approach or if you already solved that in > > another way. > > Is compile_commands.json intended only for use by IDEs? If it's not > going to break anything, then this seems reasonable to me. I would go > ahead and submit a patch for review on bugs.webkit.org, and leave a > link to it in this thread.
At least I know also that the “scan-build” implementation in Python [1] uses the compilation database (in my experience it is easier to get running than Clang's, and is what I have used in the past to run the Clang analyzer on the WebKit code). I *think* that many tools like scan-build should work equally well regardless of whether the compilation database contains the unifies or the individual sources. But of course it's difficult to predict what every other tool will do, so it's impossible to be sure :-\ Didn't we have a CMake flag that disabled unified sources completely? Maybe people who want to use “compile_commands.json” with IDEs can just use that… Cheers, -- Adrián 🎩
pgpxehPRGlYiz.pgp
Description: PGP signature
_______________________________________________ webkit-dev mailing list [email protected] https://lists.webkit.org/mailman/listinfo/webkit-dev
