Great question! I’ve messed around with Emma’s coverage files before, so let me walk you through what I know about the coverage.ec format—since it’s a binary file, you can’t just open it in a text editor and make sense of it right away.

High-Level Background #

coverage.ec is Emma’s compact binary format for storing raw coverage data. It’s designed to be small and fast to write/read, which is why it doesn’t use plaintext. The file is made up of a fixed header followed by a series of typed data records that track coverage for classes, methods, and (if enabled) individual lines of code.

Detailed Breakdown of the Format #

Let’s break it down piece by piece:

• File Header
  • Starts with a magic string: EMMA-COV (7 ASCII bytes) — this is how Emma identifies the file as a valid coverage data file.
  • Next is a 4-byte integer representing the Emma version (e.g., Emma 2.1 uses 0x00000201).
  • Then a bitmask flag (4 bytes) that tells you what kind of data is included (like whether line-level coverage is present, or if there’s a timestamp included).
• Data Records
  The rest of the file is a sequence of records, each starting with a 1-byte type code that tells you what kind of record it is. The most common ones are:
  • Class Record: Marks the start of coverage data for a specific class. Includes the fully qualified class name (UTF-8 encoded string) and a hash of the class file (to ensure it matches the compiled code Emma instrumented).
  • Method Record: Tied to a parent class, this includes the method’s signature (UTF-8), its start/end line numbers, and execution counts for each code block within the method.
  • Line Coverage Record: Only present if line-level coverage was enabled. Stores execution counts for individual lines, mapped to the corresponding class and method.

How to Parse It for Your Custom Report #

You don’t have to reinvent the wheel here—Emma’s own source code has all the parsing logic you need. If you want to build a custom parser, check out these core classes:

• com.vladium.emma.data.CoverageDataReader: This is the class Emma uses to read coverage.ec files and parse them into an in-memory data structure.
• com.vladium.emma.data.CoverageData: This holds the parsed coverage data, with methods to access class, method, and line-level coverage counts.

You can either include Emma’s jar in your project and reuse these classes directly, or use the source code as a reference to build your own parser from scratch.

Quick Debugging Tip #

If you want to see the raw coverage data in plaintext to cross-reference with the binary file, use Emma’s built-in report tool to convert it:

java -cp emma.jar com.vladium.emma.report report -r txt -in coverage.ec -out coverage.txt

The generated coverage.txt will show you all the coverage details in human-readable format, which can help you map the binary records to actual coverage data as you build your custom parser.

内容的提问来源于stack exchange，提问作者Feri