[Openmp-commits] [PATCH] D57615: [AST][OpenMP] OpenMP master Construct contains Structured block

Roman Lebedev via Phabricator via Openmp-commits openmp-commits at lists.llvm.org
Fri Feb 1 14:11:50 PST 2019


lebedev.ri added a comment.

In D57615#1381461 <https://reviews.llvm.org/D57615#1381461>, @ABataev wrote:

> In D57615#1381443 <https://reviews.llvm.org/D57615#1381443>, @lebedev.ri wrote:
>
> > Let's instead solve the problem of structured block not having an AST representation.
>
>
> Again, this is not a problem for the compiler. If you want this idiom for the analyzer, you need to teach the analyzer to handle the AST nodes in accordance with the OpenMP standard.


I'm sorry, but //to me// that just sounds completely wrong.

It sounds remotely like the gcc argument of keeping everything in
the least modular fashion just so no one does anything bad with it.

OpenMP spec gives a pretty good idea when something is an 'structured block'.
I'm specifically looking for these OpenMP structured blocks.
Are you telling me that i should forget everything i know about the 'goals' of clang AST,
(and i do believe it is reasonable to expect to have an `OMPStructuredBlock` opaque entry)
and reimplement the 'OpenMP structured blocks' finding in whatever code i'm writing,
instead of doing the right thing, and teaching the AST about it?

Obviously, said teaching should not have any negative impact on the existing clang OpenMP code.
Obviously, everyone's time is limited, and i won't be able to convince anyone to do that //for// me :)

I can see just how **quickly** i have derailed this disscussion, in less than 12h.
I guess i should stop commenting, and follow "words are cheap, show the code".


Repository:
  rC Clang

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D57615/new/

https://reviews.llvm.org/D57615





More information about the Openmp-commits mailing list