Support extracting logs from job URLs that have ?pr=123 query string

[huonw]

This makes the extract-from-URL job ID deduction for roc logs 'https://...' work when the URL has a query string or other trailing data.

This occurs in practice when clicking through to a job from a pull request's UI, e.g. clicking a job on runs-on/runs-on#245 takes one to a URL like https://github.com/runs-on/runs-on/actions/runs/13674250460/job/38231106711?pr=245, note the ?pr=245, and thus a "simple" copy-paste of those URLs from the UI will have the query strings.

The previous version of extractJobID would extract a job ID like 38231106711?pr=245 (retaining the query string) which didn't work with the rest of the code. This new version uses net/url to ensure the code is working with the URL's path alone, ignoring any extraneous data like ?queries or #fragments.