On Oct 18, 2010, at 7:47 AM, Tor Arne Vestbø wrote: > On 15.10.10 22.12, Argyrios Kyrtzidis wrote: >> On Oct 15, 2010, at 12:04 PM, Tor Arne Vestbø wrote: >> >>> On 15.10.10 20.57, Argyrios Kyrtzidis wrote: >>>> This is less about supporting directory-style precompiled headers >>>> and more about supporting Qt's build system. >>> >>> Well, yes and no. I've successfully compiled Qt with Clang using >>> the solution you outline, but I figured I'd give it a shot at >>> fixing this upstream so other people could benefit from it. >>> >>> If it's too hacky (as I admit, it does not do what GCC does, which >>> is actually parse each header found in the PCH directory and choose >>> the first matching one for the input language) I'm fine with that >>> as well -- as you say we have a solution for Qt. >> >> I don't have a strong opinion either way but I lean towards not >> adding it because it's Qt specific. Deferring to Daniel since the >> driver is his department. > > True, it's relying on the user naming their pch files after the input > language, like Qt does. > > If we don't want this patch in, would it make sense to warn the user and > then ignore the directory-pch (and fall back to -include to pick up the > real header) instead of "error: unable to read PCH file: 'Is a directory'"?
Sounds good! -Argyrios > > Tor Arne > > _______________________________________________ > cfe-commits mailing list > [email protected] > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
