CuriousGeorgiy added a comment.
@NoQ thank you for the feedback too!
> So if you want to support non-standard functions, and you know that they have
> the same portability issues, totally go for it!
AFAIC, the functions I listed are very exotic, so I would rather skip them.
> Probably a separate check would be better. The consequences of `malloc(0)`
> are likely to be much more dire than consequences of `printf("%p", 0)`, so
> people may want to enable/disable them separately.
I moved it to a separate checker, but I'm not sure about the naming, any
suggestions? BTW, I guess it should be `printf("%p", NULL)`, since as
@MitalAshok pointed out, we should only consider pointer type values.
> The run-lines are mostly self-explanatory. Just run it through the grep
> command in the other run-line. It filters out all the non-transferable stuff.
Sorry, my bad, I didn't notice this, updated the plist correctly now.
> This really doesn't accomplish anything, `assume()` already does a lot more
> than that. Just rely on `assume()`, or make it a completely path-insensitive
> check if you want to stick to literal constants.
Dropped this check, thanks for point this out!
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D154838/new/
https://reviews.llvm.org/D154838
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
