render readable ranges in look-ahead blocker diagnostics

[notatallshaw]

The Diagnostics section of a failure report explains each look-ahead blocker on its own line. Those lines are built inside the provider before the report formatter sees them, so #448's hook could not reach them and they still rendered ranges with str(). Asking for b>=2 alongside a c whose every release pins b==1.0 gave:

requires b in <VersionRange '[1.0, 1.0[AFTER_LOCALS]]'> but root has it in <VersionRange '[2, +inf)'>

They now go through the same format_range, so that line reads requires b in ==1.0 but root has it in >=2.

The decision-block line changes further. It said requires b != 1.0, negating the blocker it ran into instead of naming what the candidate asked for. It now reads requires b in ==2.0 but solution has it at 1.0, naming the decided version directly rather than as a singleton range, which has no specifier spelling of its own.

Stacked on #448.

[read-the-docs-community]

Documentation build overview

📚 nab | 🛠️ Build #33983361 | 📁 Comparing eb767a2 against latest (c273490)

  🔍 Preview build  

12 files changed · ± 12 modified

± Modified

• contributing.html
• explanation/conflicts.html
• explanation/universal.html
• how-to/install.html
• how-to/local-sources.html
• how-to/multi-index.html
• how-to/vcs.html
• how-to/workspaces.html
• reference/build-policy.html
• reference/cli.html
• and 2 more...