Codex

An extensible platform for indexing and exploring code inspired by Source Browser. This powers https://ref12.io.

Create and host your own static website to browse your C#/VB/MSBuild source code using WebAssembly.

Of course Codex allows you to browse its own source code: http://ref12.github.io/Codex

Building (in Powershell)

1. git clone https://github.com/Ref12/Codex
2. cd Codex
3. ./build.ps1 (can also call dotnet build Codex.xln direct;u)

Instructions to generate and run a website

1. Generate wasm binaries

   dotnet publish src/Codex.Web.Wasm/Codex.Web.Wasm.csproj -o bin/wasm

2. Generate Codex executable

   dotnet publish src/Codex.Application/Codex.Application.csproj -o bin/exe

3. Deploy the wasm binaries

   ./deploywasm.ps1 - deploys static site to bin/web folder by default

4. Build repo and generate binlog: ./build.ps1

5. Analyze repo and binlogs to produce analysis outputs: ./analyze.ps1

6. Create or update index ./ingest.ps1 -OutputDir bin/web/index

7. Test locally using dotnet serve

   dotnet serve -p 58226 -S --directory bin/web

Features

• Semantic analysis of C#, VB, and MSBuild
• Syntax coloring for C#, VB, MSBuild, XML
• Go To Definition (click on a reference)
• Find All References (click on a definition)
• Project Explorer - in any document click on the Project link at the bottom
• Document Outline - for a document click on the button in top right to display types and members in the current file
• Clicking on the partial keyword will display a list of all files where this type is declared
• MSBuild files (.csproj etc) have hyperlinks

Conceptual design

Indexing happens in two phases: Analysis and Ingestion. During analysis, C#/VB source files are analyzed using Roslyn and files are generated containing the extracted semantic information. These files are typically packaged into a zip file and ingested elsewhere. During ingestion, the analysis files are loaded an their information is added to a custom Lucene database. The database is used to satisfy all queries. The database format allows it to be queried in a serverless manner using WASM but this is untenable for larger codebases or aggregations of codebases so https://ref12.io uses a server based approach.

Indexing is incremental. Many repos can be analyzed and ingested at different times.

Limitations and known issues

• Analysis requires a successful build of projects or successful extraction of compilation using Roslyn MSBuildWorkspace.

License

This project is licensed with the MIT license.