[github] update API to support rerun-failed-jobs

[chantra]

Cherry-picked the rerun-failed-jobs action from an up-to-date spec. I did not include the enable_debug_logging parameter has it was not currently present with the existing re_run_workflow endpoint.

re-run on failed job doc: https://docs.github.com/en/rest/actions/workflow-runs?apiVersion=2022-11-28#re-run-failed-jobs-from-a-workflow-run
Generated diff after make github: https://gist.github.com/chantra/c30751e00fe6034a4ace4ac73646afc6#file-gist-diff-L702

[chantra]

I did not include the generated libs as for some reason my toochain still introduce those &var.to_string() changes.

Also, I am not able to actually sync the spec as a whole currently so took this shortcut.

[augustuswm]

Cool thanks! I generally pull and build these changes to check them, so no need to include the generated files.

You are correct that the generator won't build against the current API spec, for a number of reasons. That is something I'm working on. The &var.to_string() lines come from how the generator is implemented. They can be cleaned up with clippy, but that takes a long time to run.

[chantra]

That is something I'm working on.

Great. I gave it a look, but it was quite complicated 😁

The &var.to_string() lines come from how the generator is implemented. They can be cleaned up with clippy, but that takes a long time to run.

I believe we can nip it in the bud here:

third-party-api-clients/generator/src/template.rs

Lines 148 to 153 in 5480bb8

	"crate::progenitor_support::encode_path(&{}_.to_string()),",
	to_snake_case(n)
	));
	} else {
	out.push_str(&format!(
	"crate::progenitor_support::encode_path(&{}.to_string()),",

[chantra]

@augustuswm let me know if there is anything that looks off in this PR. Happy to iterate.

[augustuswm]

Looks good. Merging this and I'll get a release put together.