This is a tool for figuring out if a given piece of text contains characters that will become "tofu" on mobile platforms (specifically iOS and Android).
"Tofu" refers to the blank white square (□) sometimes shown in place of an unsupported character.
The input text is split up into UTF-16 Unicode code units, and is compared against data compiled by the CodePointCoverage project.
Coverage is determined by inspecting the font files that are bundled with the
OS. If there is at least one font providing a glyph (more specifically, with an
entry in the cmap table) for a given codepoint, that codepoint is considered
"covered".
For iOS, the fonts are taken from the iOS Simulator runtimes bundled with various versions of Xcode. For Android, the fonts are taken from the Android Emulator system images downloadable with the Android SDK.
Note that this methodology is not completely accurate. Known issues:
- Codepoints need not have their own glyphs to be considered "supported"; see e.g. variation selectors
- Codepoint-level analysis cannot account for support for multi-codepoint grapheme clusters like emoji ZWJ sequences
- Especially on Android, vendors may bundle different fonts. Even "vanilla" Android may have different fonts from the SDK (example)
If you find any inaccuracies, please feel free to open an issue.