Count a release published exactly at since, and stop early despite a draft

[Thayorns]

The release count in `total_releases_published` compared with a strict `>`, while the early-exit check compared with a strict `<`. A release published at exactly the `since` boundary satisfied neither, so it was dropped from the count without stopping the scan. Judges call this method with a boundary computed from a previous run, making that instant a likely value to land on.

The two checks also disagreed about drafts: the count skips them via `isDraft`, but the exit check only looked at `publishedAt`, which is null for a draft. A single draft anywhere on a page made the exit check false, so a repository with a draft near the top of its release list scanned its entire history one page at a time.

This makes the count inclusive on the boundary (`>= since`) and has the exit check treat a draft the same way the count does, so a draft no longer defeats the short-circuit.

Closes #662

[Thayorns]

@yegor256 please take a look, happy to clarify anything.

[yegor256]

This touches the same total_releases_published boundary condition as #671, via a different approach — only one should be merged. Assigning to @yegor256 to reconcile.

---

Generated by Claude Code

[Thayorns]

@yegor256 take a look please