GenRegexNamedBlocks Tool

Overview

This tool generates the named Unicode blocks for RegexCharClass.cs based on the Unicode Character Database (UCD) Blocks.txt file. Named blocks allow regex patterns to match characters in specific Unicode blocks using syntax like \p{IsBasicLatin} or \p{IsGreek}.

The current implementation is based on Unicode 17.0.

Usage

To update the named blocks when a new Unicode version is released:

1. Download the Blocks.txt file from the Unicode Consortium:

   Copy

   https://www.unicode.org/Public/UCD/latest/ucd/Blocks.txt
   

2. Run the tool from this directory:

   bashCopy

   dotnet run -- <path-to-Blocks.txt> ../../src/System/Text/RegularExpressions/RegexCharClass.Tables.cs
   

3. The tool will generate the RegexCharClass.Tables.cs file with all named blocks

4. Update tests in RegexCharacterSetTests.cs to include tests for new blocks if needed

5. Build and test to ensure all tests pass

Notes

• The tool automatically excludes:

  • Blocks outside the Basic Multilingual Plane (BMP) (code points >= U+10000)
  • Surrogate blocks (U+D800-U+DFFF)
  • Private Use Area blocks

• Block names are converted to "Is" + alphanumeric characters + hyphens (e.g., "Greek and Coptic" becomes "IsGreekandCoptic")

• The tool sorts blocks alphabetically by name for consistent output

• For backward compatibility, some aliases like "IsGreek" (alias for "IsGreekandCoptic") should be manually maintained

See Also

• Unicode Character Database
• Unicode Block Names
• Related tool: src/libraries/System.Text.Encodings.Web/tools/GenUnicodeRanges/