Build portaudiocpp with cmake failed

[donarturo11]

Describe the bug
Trying to create cmake build directory for portaudiocpp gets message:

CMake Error at cmake/modules/FindPortAudio.cmake:20 (message):
  PortAudio_FOUND but not target PortAudio::portaudio

To Reproduce
Steps to reproduce the behavior. Include code if applicable.

1. Install portaudio with cmake in any directory

cmake -S portaudio -B portaudio-build -DCMAKE_INSTALL_PREFIX=${PORTAUDIO_PATH}
cmake --build portaudio-build --config Release
cmake --install portaudio-build --config Release

2. Configure building with cmake portaudiocpp

cmake -S portaudio/bindings/cpp -B portaudiocpp-build -DCMAKE_PREFIX_PATH=${PORTAUDIO_PATH} 

Expected behavior
Configure build directory with no errors.

Actual behavior
CMake throws error

CMake Error at cmake/modules/FindPortAudio.cmake:20 (message):
  PortAudio_FOUND but not target PortAudio::portaudio

Desktop (please complete the following information):

• OS: Windows
• OS Version: 10
• PortAudio version: current #2fa0219
• Host API : any

Proposed solution
Unify target names cases.

Note: PortAudio is a community supported project. If you have a solution, please create a Pull Request for us to consider.

[RossBencina]

Our CI doesn't fail because our CI doesn't use portaudiocpp. But this does suggest a regression since presumably portaudiocpp used to build.

@donarturo11 would you please check PortAudio v19.7 release and confirm that it does not have this problem?

[RossBencina]

It's not failing CI because portaudiocpp is not tested by CI.

My concern now is that there is a regression in the PortAudio library CMakeLists.txt, and that maybe we should undo the library library name case change in PortAudio -> portaudio in the main PortAudio CMakeLists.txt.

@dechamps any thoughts?

[donarturo11]

I think too that undoing change case in main CMakeList will be better solution than mine.

[donarturo11]

@donarturo11 would you please check PortAudio v19.7 release and confirm that it does not have this problem?

Yes sure. So, preparing and building portaudiocpp using source from main branch and stable version of portaudio are successful.

[RossBencina]

preparing and building portaudiocpp using source from main branch and stable version of portaudio are successful.

@donarturo11 thanks. So as it stands, am I correct to say that there is a regression in our main PortAudio CMakeLists.txt ?

[RossBencina]

Seems that this was part of @Be-ing's big rewrite (which I'm now strongly thinking we should just roll-back since there is no maintainer and it has been shown to have multiple issues):

242a024#diff-1e7de1ae2d059d21e1dd75d5812d5a34b0222cef273b7c3a2af62eb747f9d20aR12

old add_library(portaudio:

242a024#diff-1e7de1ae2d059d21e1dd75d5812d5a34b0222cef273b7c3a2af62eb747f9d20aL374

old add_library(portaudio_static:

242a024#diff-1e7de1ae2d059d21e1dd75d5812d5a34b0222cef273b7c3a2af62eb747f9d20aL383

new: add_library(PortAudio:

242a024#diff-1e7de1ae2d059d21e1dd75d5812d5a34b0222cef273b7c3a2af62eb747f9d20aR12

[donarturo11]

@RossBencina Yes. In main CMakeLists.txt is a regression. Just undo cases in target names and portaudiocpp will built with no problem.

[RossBencina]

@donarturo11 Could you please close the old PR and create a new PR with our agreed fix?

[donarturo11]

@RossBencina I'm sorry for huge delay, but I had a lot of work at home. I closed old PR and I'll create new soon as possible.

[donarturo11]

So @RossBencina , could You now review a new PR, please? I hope that one will be more satysfying.