iains added a comment. In D134267#3869520 <https://reviews.llvm.org/D134267#3869520>, @tahonermann wrote:
>> In a pre-scanned world, the build system does know the info for each source >> file (published and dependent modules) [which ought to dispel some of the >> concerns raised about not knowing about possible outputs for >> implementation/interface cases]. >> >> In a discovery world, the interface to the build system carries all of this >> traffic anyway so that the command line would only be providing pre-set data >> for that. > > This does not cover all build systems. For example, Coverity relies on > observing the compiler invocations performed by another build system and > relies on interpreting the command lines of those invocations in order to > identify all inputs and outputs (and Coverity does require awareness of > outputs as well as inputs). Other tools that operate on a compilation > database or monitors like Build EAR <https://github.com/rizsotto/Bear> have > similar requirements. I did not intend to derail the discussion of how the command line option should be spelled - but for clarity - I continue to expect that we will need to support **both** styles of build system, and the user will need to choose what is appropriate for their workflow - nothing in the proposals here impacts on that, right? CHANGES SINCE LAST ACTION https://reviews.llvm.org/D134267/new/ https://reviews.llvm.org/D134267 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
