[flang-commits] [mlir] [openmp] [flang] [Flang][OpenMP] Initial mapping of Fortran pointers and allocatables for target devices (PR #71766)

[via flang-commits]

agozillon wrote:

> I will preface this by saying that the approach you've employed here looks pretty great. I like the documentation, the passes which handle some of the mapping detail after lowering, the testing, and the dialect operations changes.

Thank you very much! I am very excited to see the OpenACC implementation, I enjoyed the documentation very much, it showcased a lot of concepts and possibilities I wasn't aware of. Hopefully we'll be able to adopt that in OpenMP as well at some point in the future. 
 
> So no need to let my question block this PR. However, there is still something unintuitive here and I cannot tell why it is done this way - you may have answered for others before though. Thank you again for your work on this!

I'm happy to delay the PR if you believe it's something worth addressing. I've tried to answer your question, hopefully I understood it well enough to give a reasonable enough reply. And thank you for your review time, comments, suggestions and help, the PR wouldn't be what it is without your input. 


https://github.com/llvm/llvm-project/pull/71766