add cfitsio/3.470

[SpaceIm]

Specify library name and version: cfitsio/3.470

• I've read the guidelines for contributing.
• I've followed the PEP8 style guides for Python code in the recipes.
• I've used the latest Conan client version.
• I've tried at least one configuration locally with the
  conan-center hook activated.

[conan-center-bot]

All green in build 1 (367546acf7c3efd06481576ae73c29ef2b56f23f)! 😊

• cfitsio/3.470: Generated 82 packages. All logs here

[conan-center-bot]

All green in build 2 (c07f7980b843caedcd5f51dc255963de264293e3)! 😊

• cfitsio/3.470: Generated 82 packages. All logs here

[conan-center-bot]

All green in build 3 (f52816410df27fe287f9e828f3c41591db31a86c)! 😊

• cfitsio/3.470: Generated 82 packages. All logs here

[madebr]

Why did you create a new cmake build script?
The tarball contains a autotools and cmake build script.

[uilianries]

Use the CMakeLists.txt provided by the author. Even other pkg managers are using it.

[SpaceIm]

I'm not confortable enough with autoconf and its helper class in conan to provide a recipe, therefore I prefer to let someone else create a recipe for cfitsio ;)
There will be several additional patches if you only rely on CMakeLists: cfitsio CMakeLists lacks several features compared to autoconf files (shared memory, simd, few definitions to inject, install only public headers, bzip2 support, proper support of remote file driver).

[madebr]

@SpaceIm What I have done in the past is fork the project's repo (or use the release as a base),
fix some problems and send my patches upstream.
It seems they have an e-mail address listed on the homepage.
You can then re-use that patch here in CCI.

[SpaceIm]

@madebr @uilianries is it acceptable with this patch on CMakeLists?

And yes, I could send an email to library authors.

[madebr]

@madebr @uilianries is it acceptable with this patch on CMakeLists?

And yes, I could send an email to library authors.

I ❤️ it! 😄

[conan-center-bot]

All green in build 4 (2dba85a33fc7ebd30ab3647e0a24b57095ecb2f6)! 😊

• cfitsio/3.470: Generated 82 packages. All logs here

[SpaceIm]

Need to check compiler.threads for MinGW in order to require pthreads4w if win32 threads (also update CMakeLists to link to link to CONAN_PKG::pthreads4w if CMAKE_USE_WIN32_THREADS_INIT True).

[uilianries]

Only few improvements, but looks good. Please, send your patch to author.

[conan-center-bot]

All green in build 5 (da3c01f9e1832143bed2e78e8e48b3140960d333)! 😊

• cfitsio/3.470: Generated 82 packages. All logs here

[conan-center-bot]

All green in build 6 (d10182fa5a178e6d285fe7957d74aa71476d2e9c)! 😊

• cfitsio/3.470: Generated 82 packages. All logs here

[jgsogo]

Awesome work on the build system. Did you finally send the patch to the library authors? Did they answer anything? I'm sure they can take advantage of the improvements here... and also their investigators would love to run just a conan install ... within their project to have the dependencies they want working in their project.

Great!

[SpaceIm]

Not yet, the patch requires more love to be submitted to library authors. They can't rely on CONAN_PKG in their own CMakeLists and two options should be added in order to conditionnaly build test and executable.

[jgsogo]

When I add a recipe to CCI, I always use the cmake_find_package generator, it provides proper target names to avoid using the CONAN_PKG:: one.

I really hope that all these efforts will eventually arrive upstream and we are contributing to making the original projects better too.

[SpaceIm]

I don't trust cmake_find_package with current conan state, since components are not implemented (maybe I'm wrong).

[jgsogo]

Components are not implemented from the consumer perspective. I would really like to know what you don't trust from the current state of that generator. Components will add only granularity, but the functionality within each component/target will be essentially the same.

Feel free to contact me slack, open a Github issue or whatever works for you.

Thanks!