[BUG] Diff/search/replace alters HTML entities (unexpected code changes)

[hannesrudolph]

Type

Bug

Problem / Value

When applying changes, encoded characters like &, <, and > are sometimes changed or removed. This leads to unintended edits and can break code or markup, causing confusion and rework.

Context

Users applying diffs or performing search/replace see encoded characters (HTML entities) modified in ways they didn’t intend. This has been raised multiple times and earlier fix attempts did not land, so the behavior persists. A consistent, predictable outcome is needed where encoded text is preserved exactly as written.

Constraints/Preferences

• Preserve encoded characters exactly as authored; avoid automatic conversion between entities and literal characters
• Ensure behavior is consistent across single-file and multi-file change flows
• Add regression coverage to prevent recurrences

[roomote]

I'm exploring a potential fix for the HTML entity preservation issue and will open a PR to share progress.

[roomote]

Opened a PR with my attempt at addressing this issue: #8071

The fix removes HTML entity unescaping from the diff and file operation tools to preserve encoded characters exactly as written. Feedback welcome!

[daniel-lxs]

I think to solve this properly we need to add a setting that enables or disables automatic HTML entity escaping.