On Mon, Jul 15, 2013 at 6:28 PM, Hans Wennborg <[email protected]> wrote: > On Mon, Jul 15, 2013 at 6:11 PM, David Blaikie <[email protected]> wrote: >> (not claiming ownership of this CR, but a few drive-by comments) >> >> You could consider using a StringSwitch in Driver::ParseCCCMode. > > Done. It's a little annoying that I need an invalid value for the > default case, but on the other hand the StringSwitch is nice.
Fair enough - not a strong opinion given the drawback, so whichever you prefer. (including Reid's 3rd variation) > >> Is there any reason you need to remove support for "clangxx" (that's >> causing you to need to update those hexagon tests?)? > > I'm not, I'm switching those hexagon tests over to use clangxx instead > of -ccc-cxx which I'm removing. Ah, right, misread. Alternatively you could migrate to the --driver-flavor (just checking my understanding)? I suppose your change makes (at least as much) sense, then. > >> Would you mind updating immediate-options.c to use FileCheck rather >> than multiple invocations with grep? > > Sure. I'd like to do that in a separate patch, though. Sure thing. > Thanks for the comments! New patch attached. > > - Hans > > >> On Mon, Jul 15, 2013 at 4:17 PM, Hans Wennborg <[email protected]> wrote: >>> Hi all, >>> >>> This patch is a follow-up to the discussion in Reid's cl.exe >>> compatible driver proposal [1]. >>> >>> Clang currently looks at argv[0] to switch between running in >>> gcc/g++/cpp mode. This mode is represented in Driver by CCCIsCXX and >>> CCCIsCPP. Since there is no overlap between the three modes, I have >>> turned those two variables into an enum instead. The plan is to later >>> extend this enum with a "cl.exe mode". >>> >>> The patch also adds a new command line option, -ccc-mode, to set the >>> mode. This replaces the current -ccc-cxx option (and also makes it >>> easier to test the cpp mode). This option is special: because the mode >>> can affect option parsing (the cl.exe mode will add new options), the >>> -ccc-mode needs to be parsed early. >>> >>> Please take a look! >>> >>> Thanks, >>> Hans >>> >>> [1]. http://lists.cs.uiuc.edu/pipermail/cfe-dev/2013-June/030439.html >>> >>> _______________________________________________ >>> 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
