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

Tue Oct 18 00:11:13 PDT 2022

ChuanqiXu added a comment.

In D134267#3864248 <https://reviews.llvm.org/D134267#3864248>, @iains wrote:

> I am also OK with doing this in two steps (first in the driver with this patch and then by updating the FE to allow the two outputs from one invocation - my draft patch series).

For your draft patches, I have only one concern in the high level: if we will/should restrict the module declarations appear in the filenames with special suffixes. Since both @rsmith and MSVC made the same decision. I am not sure if they have special reasons to do so. Maybe I need to search the discussion in SG15 mailing lists. And another related concern (maybe concern is not a good word here) is, if your patches landed, many existing codes may need to be removed. Otherwise, it'll be redundant codes. But this might not be a blocking issue though.

> BTW: I did mean to ask before .,, did you consider this (existing) command syntax?
>
> `-fmodule-file=[<name>=]<file>`
>
> and see if it works for your case? (it seems that it should to be consistent).

Oh, do you mean we should use `module-file` name since we've used `-fmodule-file` option? Good point. Yeah, it looks like a pity to have 2 terms to describe the same thing. Personally I don't have strong feeling for the option name. If @dblaikie has no other comments on this, I'll follow your suggestion.

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

https://reviews.llvm.org/D134267