[PATCH] D134267: [C++] [Modules] Support one phase compilation model for named modules

[Iain Sandoe via Phabricator via cfe-commits]

iains added a comment.

In D134267#3870064 <https://reviews.llvm.org/D134267#3870064>, @ChuanqiXu wrote:

> I grepped `options.td` and got (incomplete) list for options to take a output name:
>
>   # -o and its alias
>   -o
>   -object_file_name=
>   --output=
>   
>   /Fa (windows for assembly output filename)
>   /Fe (windows for output executable file name)
>   /Fi (windows for preprocessed output filename)
>   /Fo (Windows for object file)
>   
>   -dependency-dot (for DOT-formatted header dependencies)
>   -dependency-file (to write dependency output to)
>   
>   -header-include-file (Filename to write header include output to)
>   
>   -opt-record-file (Filename to use for YAML optimization record output)
>   
>   -split-dwarf-output (Filename to use for split dwarf debug info output)
>   
>   -stack-usage-file (to write stack usage output to)
>   -coverage-data-file (Emit coverage data to this filename)
>   -coverage-notes-file (Emit coverage notes to this filename)
>
> And it looks like the `-file` appears a lot. So may be the suggestion (`-fc++-module-file-output`) may be better. And for the default location, I feel like my explanation above makes sense. If the end user wants to produce .pcm files, they can use `--precompile` just like what they do with `-c` to get the object files. This only matters with end users since the build systems should/would chose other positions.

OK. I guess the idea about `-fmodule-file=<name>=filename` was that, because the FE will not try to read `filename` (for module-generation cases) we could use it to describe the output file.  However, it seems that might be too complex... so `-fc++-module-file-output` seems OK to me.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D134267/new/

https://reviews.llvm.org/D134267