Shaderc new port,update vulkan-<all>, glslang, spirv-tools, spirv-headers

[christophecvr]

Description

See: https://trac.macports.org/ticket/71081
shaderc: Added Newport
vulkan-<all>: update to 1.3.296.0
glslang: update to 1.3.296.0
spirv-tools: update to 1.3.296.0
spirv-headers: update to 1.3.296.0

Set where it was not done yet the python used version to 3.12
Fixes: https://trac.macports.org/ticket/71081

Type(s)

• bugfix
• enhancement
• security fix

Tested on

macOS 10.15
Xcode 11.5/ Command Line Tools 11.5

Verification

Have you

• followed our Commit Message Guidelines?
• squashed and minimized your commits?
• checked that there aren't other open pull requests for the same change?
• referenced existing tickets on Trac with full URL in commit message?
• checked your Portfile with port lint --nitpick?
• tried existing tests with sudo port test?
• tried a full install with sudo port -vst install?
• tested basic functionality of all binary files?
• checked that the Portfile's most important variants haven't been broken?

[macportsbot]

Notifying maintainers:
@judaew for port glslang, spirv-headers, spirv-tools, vulkan-headers, vulkan-loader, vulkan-tools, vulkan-utility-libraries, vulkan-validationlayers, vulkan-volk.

[christophecvr]

@judaew If You wish I can remove this pull request again and remake one with the last commit squashed in.
Then we have just a pull request consisting of 4 commits. :

• upgrade spirv-tools, spirv-headers
• upgrade glslang
• upgrade vulkan-<all>
• Adding new port shaderc

Squashing it all in one single commit is to much I find

But they well needs to be upgraded all togheter

[judaew]

Looks got to me, I see a lot of work has been done. Thank you! I'll try to build it on my laptop tonight.

[judaew]

Squashing it all in one single commit is to much I find

I agree with you.

It's great that you updated glslang! I couldn't figure out what was wrong with it.

[christophecvr]

Now that find out the full functionality and goal of building with the trace option, I found a lot more which is not 100% ok .
It actually never was

Squashing it all in one single commit is to much I find

I agree with you.

It's great that you updated glslang! I couldn't figure out what was wrong with it.

Yes glslang took me a lot of time before i found out how to best upgrade it.

However thanks to all this trouble with the fact that the intial checks failed. I discovered finally the real use of -t (trace) option into a build. I also dscovered a bug (buggy use op GLOB_RECURSE into the vulkan-loader port.
To create the vulkan.framework it does a search of all possible needed header files. Unfortunately this procedure will start a scan of all files into /opt/local/include. When building with trace after two hours it is still stuck during configure phase.
Googling on the web i did found that here it is a wrong use of the GLOB_RECURSE.
The error is :
loader/CMakeLists.txt line number 463
file(GLOB_RECURSE CONFIGURE_DEPENDS FRAMEWORK_HEADERS ${VulkanHeaders_INCLUDE_DIRS})
the ${VulkanHeaders_INCLUDE_DIRS} is just /opt/local/include

I'm currently working on it to solve this issue. Now it search into a massive amount of places.

[christophecvr]

p.s. without the -t option it does build ok but it well hangs for about 25 minutes during config of vulkan-loader port.

[christophecvr]

So try it all out already by you. In the mean time I'm preparing a new push request with a couple of improvements concerning not set depends-build_append ports and or required tools. Also I will include a patch for the vulkan-loader port.

[christophecvr]

Ok I will close this pull request, and then send another complete clean one.

[judaew]

I checked again this PR, everything looks is goods. Thank you so much!

Ok I will close this pull request, and then send another complete clean one.

Then I will merge it immediately after CI checking. I could to merge this PR, but you beat me to it and closed.