EML2PDF

EML2PDF is a .NET CLI utility that converts an email file (.eml) into a PDF.

What the solution currently does

At runtime, the app:

1. Loads settings from appsettings.json located next to the executable.
2. Configures Serilog with:
   • rolling local file logs (logs/log-*.txt, daily rolling, last 7 files retained)
   • Seq sink (address and optional API key from config)
   • logs are flushed via Log.CloseAndFlushAsync() in a finally block at the end of every run
3. Reads the input path either:
   • from the first CLI argument, or
   • from console prompt input.
4. Validates that the path is an existing file or directory:
   • If it is a single .eml file, that file is processed.
   • If it is a directory, every top-level *.eml file in it is processed (non-recursive). A failure on one file is logged and counted but does not stop the rest of the batch.
5. For each file, searches for PDF attachments at the deepest nesting level:
   • If any PDF attachments are found, all of them are extracted and saved to the same folder, using each attachment's original filename.
   • If no PDF attachments are found, the deepest nested .eml body is rendered via headless Chromium (PuppeteerSharp) and exported to PDF as <originalName>.eml.pdf.
6. Writes the output in the same folder as input using <originalName>.eml.pdf.
7. If successful, renames the source .eml to <yyyyMMddHHmmss> - <name>.eml.bak in the same folder (UTC timestamp, no spaces, seconds precision).
8. When the run finishes, writes a summary log entry with the input type (file/directory), total elapsed processing time, and the number of files processed, succeeded, and failed.

Usage

Download the latest release package. It contains the executable and appsettings.json.

Run one of:

• Interactive mode:

  EML2PDF.exe

• Direct argument mode (single file):

  EML2PDF.exe "C:\path\to\mail.eml"

• Direct argument mode (directory — processes every top-level .eml file):

  EML2PDF.exe "C:\path\to\folder"

Configuration

appsettings.json keys currently used by the app:

• Seq:ServerAddress: Seq endpoint URL.
• Seq:ApiKey: API key used to authenticate with Seq. Leave empty to send events without authentication.

Example:

{
  "Seq": {
	 "ServerAddress": "http://localhost:5341/",
	 "ApiKey": ""
  }
}

Output and Exit Codes

• Success output: a line prefixed with RET-OUTPUT: <pdfPath>.
• When multiple PDF attachments are extracted, one RET-OUTPUT: line is printed per file.
• On completion a summary line is printed to the console and logged, e.g. Done. Processed 5 file(s) in 00:00:12.3456789. Succeeded: 4, Failed: 1.
• Exit code 0: all processed files succeeded (or their output already existed).
• Exit code 1: invalid input path, at least one file failed, or a runtime error occurred.

Known resilience gaps (current implementation)

The following issues are currently present in code and should be considered known limitations:

1. Startup errors are not globally guarded. Fixed: all of Main is wrapped in a single try/catch(Exception)/finally; CloseAndFlushAsync is called unconditionally in the finally block.
2. Seq sink is always enabled when logger is built.
   • Bad/missing Seq endpoint can cause logging pipeline failures depending on environment/config.
3. BrowserFetcher().DownloadAsync() runs for every conversion.
   • Network or filesystem failures can break processing; cold-start latency is high.
4. Output file name is fixed to <input>.eml.pdf.
   • Existing file causes an early return instead of deterministic overwrite/versioning policy.
5. Source file cleanup (delete/move) is not wrapped in local error handling.
   • Permission/lock issues can fail after a successful conversion.
6. Backup naming has minute-level timestamp resolution.
   • Multiple runs in same minute can collide on .bak name.
7. MIME traversal/decoding assumes all nested payloads are valid.
   • Corrupt nested .eml or malformed parts can throw and abort.
8. HTML rendering does not sanitize/contain remote resources.
   • Embedded/linked assets may fail to load or behave inconsistently in headless rendering.
9. PDF page sizing uses document.body.scrollHeight as a single page height.
   • Very long messages can create oversized single-page PDFs and stress rendering.
10. No cancellation or timeout boundaries.
    • Stalls in MIME parsing, browser startup, network download, or PDF generation can hang the process.

Dependencies

• .NET 10
• MimeKit
• PuppeteerSharp (Chromium download and headless rendering)
• Serilog + File sink + Seq sink