[PATCH] D132266: [Clang][SemaChecking] move %hh and %h -Wformat warnings to -Wformat-pedantic

Mon Aug 22 07:13:25 PDT 2022

aaron.ballman added a comment.

In D132266#3739542 <https://reviews.llvm.org/D132266#3739542>, @inclyc wrote:

> In D132266#3739513 <https://reviews.llvm.org/D132266#3739513>, @aaron.ballman wrote:
>
>> Thanks for working on this @nickdesaulniers! I think we actually want to go a slightly different direction than this and disable the diagnostics entirely. Basically, we should be make sure the format specifier diagnostics are accounting for the clarifications in https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2562.pdf. So the `h` and `hh` modifiers would not even be pedantic warnings in this case.
>>
>> This should also have a release note associated with it and if you think you've completed support for N2562, the clang/www/c_status.html page should update the `Partial` markings.
>
> Sorry, I have some questions about this clarification agreement (currently working on N2562). The length modifier `hh` says in the standard that it should point to `signed char` or `unsigned char`, and if an `int` parameter is passed, why shouldn't we give such a warning? (even if it's pedantic somehow)

That's for the `fscanf` functionality, which is different than the `fprintf` functionality that's being addressed here. We should be diagnosing the `fscanf` type mismatch cases, per 7.23.6.2p12: "In the following, the type of the corresponding argument for a conversion specifier shall be a pointer to a type determined by the length modifiers, if any, or specified by the conversion specifier."

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D132266/new/

https://reviews.llvm.org/D132266