Re: Surveying interest regarding CMake
| From: | Horváth V. | Date: | Mon, 03 Jan 2022 18:33:47 +0000 |
| Subject: | Re: Surveying interest regarding CMake | ||
| References: | 1 2 3 4 5 6 7 8 9 10 | Groups: | php.internals |
| Request: | Send a blank email to internals+get-116791@lists.php.net to get a copy of this message | ||
On 2022. 01. 03. 18:17, Christoph M. Becker wrote:
Oh, that would be an issue. We can't use Cygwin builds; MinGW builds might be okayish, though. ICU ships a VS solution file (source/allinone/allinone.sln) which works fine. I don't know whether using that would be okay for Conan.No need to worry about that, it's just a way to coerce ICU's build system into behaving well on Windows. Conan uses Msys2 as a build environment for projects with *nix lock-in, which produces native binaries without the cygwin shenanigans.
Also note that we're using a patched libffi[1]; since that patch is not complete, I didn't submit it upstream, but at least this basic __vectorcall support is important for ext/ffi.That's a tough one. I see you commented on the issue regarding this: https://github.com/libffi/libffi/issues/500#issuecomment-671037484 I'm not sure what should be done in this case. I have a feeling the Conan team would not welcome such a partial feature patch. Maybe having an Artifactory server on php.net, where a custom recipe is hosted would solve this. I'm sure if the Conan team was contacted about this, then they would be glad to help.