feat(maven): Leverage index HTML for Maven Central

[zharinov]

Changes:

• When fetching from Maven Central, parse package index.html file instead of metadata
• For any error, fallback to current implementation

Context:

• Ref: Renovatebot logs continuously shows 'external-host-error'  #12576
• Ref: refactor(maven): Simplify datasource code #12589

Documentation (please check one with an [x])

• I have updated the documentation, or
• No documentation update is required

How I've tested my work (please tick one)

I have verified these changes via:

• Code inspection only, or
• Newly added/modified unit tests, or
• No unit tests but ran on a real repository, or
• Both unit tests + ran on a real repository

[zharinov]

DEBUG: http statistics (repository=renovate-testing/test-3866-is-conflicted-refactor)
"urls": {
  "https://api.github.com/graphql (POST)": 4,
  "https://api.github.com/repos/renovate-testing/test-3866-is-conflicted-refactor/commits/renovate/org.apache.lucene-lucene-queries-8.x/check-runs (GET)": 1,
  "https://api.github.com/repos/renovate-testing/test-3866-is-conflicted-refactor/commits/renovate/org.apache.lucene-lucene-queries-8.x/status (GET)": 1,
  "https://api.github.com/repos/renovate-testing/test-3866-is-conflicted-refactor/pulls (GET)": 1,
  "https://repo.maven.apache.org/maven2/org/apache/apache/23/apache-23.pom (GET)": 1,
  "https://repo.maven.apache.org/maven2/org/apache/lucene/lucene-parent/8.10.1/lucene-parent-8.10.1.pom (GET)": 1,
  "https://repo.maven.apache.org/maven2/org/apache/lucene/lucene-queries/8.10.1/lucene-queries-8.10.1.pom (GET)": 1,
  "https://repo.maven.apache.org/maven2/org/apache/lucene/lucene-queries/index.html (GET)": 1,
  "https://repo.maven.apache.org/maven2/org/apache/lucene/lucene-solr-grandparent/8.10.1/lucene-solr-grandparent-8.10.1.pom (GET)": 1
},
"hostStats": {
  "api.github.com": {"requestCount": 7, "requestAvgMs": 327, "queueAvgMs": 0},
  "repo.maven.apache.org": {
    "requestCount": 5,
    "requestAvgMs": 290,
    "queueAvgMs": 0
  }
},
"totalRequests": 12

[zharinov]

We may want to live test for Maven Central and then spread this approach to other common Maven registries

[rarkins]

Any impact on our ability to detect each version's release date?

[zharinov]

Any impact on our ability to detect each version's release date?

We're losing seconds/millisecond precision, which seems to be insufficient. However, I think we need to apply both approaches. For older releases, HTML parsing is okay. For newer ones, we still may have to make additional queres (if release is absent in HTML).

So, marking PR as draft for now.

[zharinov]

Tested with real repo

[renovate-release]

🎉 This PR is included in version 29.9.0 🎉

The release is available on:

• GitHub release
• 29.9.0

Your semantic-release bot 📦🚀

[astellingwerf]

Is the release timestamp an absolute requirement? I'm asking in the light of #5667. For example, if you look at the https://repo1.maven.org/maven2/ant/ant/, you'll see that there is no timestamp for 1.7.0 (and it's not advertised in the maven-metadata.xml) but it's also the sole version that declares the relocation of the group from "ant" to "org.apache.ant".

[rarkins]

@astellingwerf probably not necessary to be mandatory, please raise as a bug report issue