thakis added a comment. In D63648#1591469 <https://reviews.llvm.org/D63648#1591469>, @aganea wrote:
> It totally makes sense, thanks for the explanation Nico! Let's forget about > `cl` compatibility, that wasn't my initial intent. > > We always pass //relative// paths on the cmd-line, for all the reasons you've > mentioned. We also pass `-fdiagnostics-absolute-paths` hoping to fix the > Visual Studio jump-to-location issue I've mentioned above. However that > doesn't work, because we do: > > (locally, on my PC) > $ clang-cl file.cpp -Isome/relative/path/ -fdiagnostics-absolute-paths > /showIncludes -E >file_pre.cpp > > (and then, on a remote PC) > $ clang-cl file_pre.cpp -fdiagnostics-absolute-paths > (the remote stdout is then displayed on my PC) > > > `-fdiagnostics-absolute-paths` has no effect in the latter case, because the > paths are extracted from the preprocessed file, and paths can't be > absolutized anymore on the remote PC because it doesn't have the source code > (it is only a worker server). > So we end with relative paths in diagnostics, and Visual Studio can't jump > to the location of the diagnostic. > > What would you suggest? Use absolute paths along with `-fdebug-prefix-map`? I see, thanks for the details. A few more questions (sorry!): - Does fastbuild have something like distcc-pump? (In pump mode, distcc sends the raw cc files and all included headers to the server, and preprocessing happens remotely – http://manpages.ubuntu.com/manpages/bionic/man1/distcc.1.html#how%20distcc-pump%20mode%20works) I think this would address this (?) - IIRC we used to use fdiagnostics-absolute-paths in Chromium but eventually stopped because someone figured out how to make MSVC's jump-to-diag work even with relative paths (?) I'm not sure what the Right Fix is here – I can see use cases for `-E` both with relative and with absolute paths, even if `-fdiagnostics-absolute-paths` :-/ I think `-fdebug-prefix-map` doesn't affect `-E` output as far as I know. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D63648/new/ https://reviews.llvm.org/D63648 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
