Fix Windows benchmark failure: add UTF-8 encoding to all file I/O

[Copilot]

The benchmark (windows-latest) job was failing immediately after startup because generate.py and generateog.py called open() without encoding='utf-8'. On Windows, the default encoding is cp1252, which chokes on the UTF-8 content present throughout the repo.

The crash happened at module load time: _load_properties(LOCALES_FILE) is called at the top level of generate.py, and locales.properties contains flag emojis and non-ASCII locale names (🇬🇧, 中文, العربية, etc.) that cp1252 cannot decode. Content YAML files compound this with emoji in whyModernWins icons (📏, 🔒, etc.).

Changes

• generate.py — add encoding="utf-8" to all 6 open() calls (properties, YAML/JSON reads, HTML/JSON/index writes)
• generateog.py — same fix for all 3 open() calls (properties, YAML/JSON reads, SVG writes)
• benchmark.yml — fix both snippet_count calculations to match .yaml/.yml/.json files; the old -name '*.json' filter found only content/template.json, reporting "1 snippets" instead of ~112

# Before — crashes on Windows when reading locales.properties
with open(path) as f: ...

# After
with open(path, encoding="utf-8") as f: ...

Original prompt

Fix the failing GitHub Actions workflow benchmark (windows-latest)
Analyze the workflow logs, identify the root cause of the failure, and implement a fix.
Job ID: 65349663712
Job URL: https://github.com/javaevolved/javaevolved.github.io/actions/runs/22561737825/job/65349663712

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.