build: use zig 0.13.0

[hendriknielaender]

• update README.md
• github action
• use std.process.Child
• build.zig lazyPath
• new std.Progress api
• added zon file

Closes #84

[ivanstepanovftw]

And migration:

exe.addAnonymousModule("zbench", .{
  .source_file = .{ .path = "libs/zbench/zbench.zig" },
});

V

exe.root_module.addAnonymousImport("zbench", .{
  .root_source_file = b.path("libs/zbench/zbench.zig"),
});

[hendriknielaender]

@bens @FObersteiner does somebody have a smart idea to integrate the new progress api?

[FObersteiner]

@bens @FObersteiner does somebody have a smart idea to integrate the new progress api?

unfortunately I don't. Personally I haven't found a good use for this API, while at the same time it seems like a considerable amount of work to get into it. Andrew Kelley even wrote a blog post about it, but I wouldn't call this "user-facing documentation". Although, thread-safe, lock-free, and avoid contention sounds promising.

[hendriknielaender]

@FObersteiner I removed the old std.Progress, and leave the new progress api out of scope for this PR.

For the failing macOS platform test, I will have a look couldn't re-produce it on my machine.

[FObersteiner]

@FObersteiner I removed the old std.Progress, and leave the new progress api out of scope for this PR.

For the failing macOS platform test, I will have a look couldn't re-produce it on my machine.

added a zon file for convenient usage of the library in other projects.

doesn't fix the CI error ofc ^^ looks like it's coming from the platform info example. @hendriknielaender did you try to run it with a GPA instead of the fixed buffer allocator? Maybe this gives more helpful debug output.

[hendriknielaender]

@FObersteiner increased fixed buffer alloc to from 4096 to 8192 fixed the issue.
Maybe there is a smarter way to determine the fixed alloc for each platform.

[FObersteiner]

@FObersteiner increased fixed buffer alloc to from 4096 to 8192 fixed the issue. Maybe there is a smarter way to determine the fixed alloc for each platform.

right... nevertheless, even only 4k seems like overkill to parse an int

[hendriknielaender]

@FObersteiner increased fixed buffer alloc to from 4096 to 8192 fixed the issue. Maybe there is a smarter way to determine the fixed alloc for each platform.

right... nevertheless, even only 4k seems like overkill to parse an int

You are right, i created an issue to fix this even 4096 like you mentioned seems to overkill.
#88