Szelethus accepted this revision.
Szelethus added a comment.
This revision is now accepted and ready to land.

Ugh, yeah, so it has come to this. I championed my idea of granulalizing 
checkers into modeling and reporting components since what... 2018? I think the 
goal is still something to shoot for, but its time to admit that we the 
underlaying infrastructure needs to encourage that a lot more (and there is 
already so much dept in the form of these giant checkers).

I like the idea that the modeling portion and the reporting portion are in 
different checkers, but that is hard to achieve, when fatal errors by the sheer 
fact that they create sinkholes makes this practically impossible (we'd have to 
rewrite almost EVERY checker with that in mind, and also in way that it wont 
just require a branch). Even then, I'm not sure we would want that -- our 
checkers are not perfect, and yes, we sometimes do want to explicitly disable a 
non-alpha checker and everything that they do, including modeing because they 
are not working well on a project.

For how much I took a dump on checker silencing back in the day, I'm coming to 
terms with the fact that it is likely to be the most sustainable approach 
moving forward. Or, I don't know, the sheer size of these checkers really 
scream for some splitting up...

In any case, LGTM. I'm sure there are other ways around this, like creating an 
inner checker callback system inside this checker, but that sounds ridiculus, 
and its not like we can't do it after the fact.

Please grep the codebase for `apiModeling.StdCLibraryFunctionsChecker` to find 
other references to the now removed checker. Do the docs need any adjustment?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D151225/new/

https://reviews.llvm.org/D151225

_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to