[PATCH] D93224: [RFC][analyzer] Use the MacroExpansionContext for macro expansions in plists

Tue Jan 12 10:40:47 PST 2021

steakhal added a comment.

In D93224#2493515 <https://reviews.llvm.org/D93224#2493515>, @xazax.hun wrote:

> In D93224#2493434 <https://reviews.llvm.org/D93224#2493434>, @steakhal wrote:
>
>> How should I continue to get this working with CTU?
>
> We have two CTU modes. One loading the dump and the other loading from AST source file. I assume that you refer to the former?

Yes, I was referring to that.

> Could you validate that this solution works for the latter?

Sure, I suspect that will have the required PP events since the Preprocessor is parsing source code :D
I'm going to check this just to be sure.

> If we cannot make CTU fully functional, is it possible to at least make it work for the current TU (as opposed to functions inlined from another TU)?

This element of the patch stack implements exactly that.

> While it would be great to have a fully working solution, not having macro expansions in a single relatively narrow scenario might worth the trade-off to get rid of the related bugs and throw out a half-baked reimplementation of the preprocessor.

Even in the CTU case, the main TU is still parsed via the main Preprocessor - hence we will track the macro expansions accordingly.
I would definitely favor this 'option' instead of dropping this patch stack.
I mean, it really fixes a bunch of crashes and expands macros much more accurately especially for the corner-cases.

> If we really want to make it work for the CTU mode with dumps, we could dump the macro expansions to a separate file together with the AST and read it back for the error reporting.

If we chose this path, the local `getExpandedMacro` function needs to be simplified. That would help Artem's dependency issue.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D93224/new/

https://reviews.llvm.org/D93224