[BUG]: Incorrect check for cudax flag

[osayamenja]

Is this a duplicate?

• I confirmed there appear to be no duplicate issues for this bug and that I agree to the Code of Conduct

Type of Bug

Something else

Component

CCCL

Describe the bug

In cccl-config.cmake, the check for CCCL_ENABLE_UNSTABLE always returns false.

Instead, the check should be

if(DEFINED ENV{CCCL_ENABLE_UNSTABLE} AND "$ENV{CCCL_ENABLE_UNSTABLE}")
     # ....
endif()

How to Reproduce

Set CCCL_ENABLE_UNSTABLE=ON or any truth value and try to import CCCL::cudax

Expected behavior

Doing the above should include the cudax target.

Reproduction link

Operating System

N/A

nvidia-smi output

N/A

NVCC version

N/A

[jrhemstad]

@alliepiper any chance you already took care of this as part of the cleanup you're doing?

[alliepiper]

Nope, but I'll add it to the list.

[alliepiper]

This is working as intended. That is not an environment variable, but rather a cmake variable that should be defined before find_package(CCCL).

set(CCCL_ENABLE_UNSTABLE ON)
find_package(CCCL)

[osayamenja]

This is working as intended. That is not an environment variable, but rather a cmake variable that should be defined before find_package(CCCL).

set(CCCL_ENABLE_UNSTABLE ON)
find_package(CCCL)

Yes, verified that this works! I misunderstood the README, thinking it was an env variable.