[PATCH] D110663: [Driver] Support Debian multiarch style lib/clang/14.0.0/x86_64-linux-gnu runtime path and include/x86_64-linux-gnu/c++/v1 libc++ path

Tue Nov 23 10:55:59 PST 2021

lebedev.ri added a comment.

Were you able to actually reproduce the problem that lead to revert of D107799 <https://reviews.llvm.org/D107799>, or is this based on blind guesses?

In D110663#3149364 <https://reviews.llvm.org/D110663#3149364>, @MaskRay wrote:

> In D110663#3148690 <https://reviews.llvm.org/D110663#3148690>, @sylvestre.ledru wrote:
>
>> I am not sure why you are pinging this review often?! Maybe chat with people who gave you feedback before directly?!
>
> https://llvm.org/docs/CodeReview.html "If it is not urgent, the common courtesy ping rate is one week." I did it less frequently than once per week.
>
> Ping @phosek
>
>> Myself I wasted so much time because of D107799 <https://reviews.llvm.org/D107799> that I am very reluctant to enable LLVM_ENABLE_PER_TARGET_RUNTIME_DIR again.
>> The mismatch of the triple / quadruple between the system triple and what have been decided in llvm ( x86_64-linux-gnu vs x86_64-pc-linux-gnu)
>>
>>> Moreover, llvm-project has always had the attitude: "downstream projects are on their own".
>>
>> I am not sure it is a correct statement. llvm project is a bunch of people with different focus. 
>> But in general, we are trying to be nice with the ecosystem and people who rely on this software.
>
> Still correct. For courtesy I typically notify projects but many changes (especially IR and codegen changes) cannot really notify users simply because you cannot enumerate every user and
> in some IR/codegen folks's minds this would unnecessarily slow down development.
> For this case I have notified rust, which may be the most likely impacted user. I think the mostly likely impacted user was probably just Fuchsia, but I have added a hack to exclude Fuchsia.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110663/new/

https://reviews.llvm.org/D110663