jstedfast/MimeKit

MaxMimeDepth=1 doesn't set HtmlBody and TextBody

firat-plutoflume opened this issue · 3 comments

Describe the bug
We are trying to minimize our cpu and memory usage by only extracting first level of attachments/emails under an email. MaxMimeDepth option seems to designed for this, but when I pass new ParserOptions() { MaxMimeDepth = 1 } to MimeMessage.Load() it doesn't set HtmlBody and TextBody fields (although I can see them under BodyParts collection)

I believe the reason is we get MimePart instead of a TextPart here when depth=MaxMimeDepth, then it fails check in one of these places (just guessing):

Note: I'm not 100% sure if this is intentional or a 'bug', if not could you please guide how to extract only first level of attachments and have the body fields?

Platform (please complete the following information):

  • OS: Linux
  • .NET Framework: .Net Standard v2.0
  • MimeKit Version: 4.1.0

To Reproduce
Steps to reproduce the behavior:

  1. Call MimeMessage.Load(new ParserOptions() { MaxMimeDepth = 1 }, stream)
  2. Verify there are body parts with text/html and text/plain types
  3. Verify MimeMessage.TextBody and MimeMessage.HtmlBody are null

Expected behavior
I expect MimeMessage.TextBody and MimeMessage.HtmlBody to be not null when there are body parts with text/html and text/plain types

Code Snippets
If applicable, add code snippets to help explain your problem.

using (var stream = new MemoryStream(mimeBytes))
{
    var res = MimeMessage.Load(new ParserOptions() { MaxMimeDepth = 1 }, stream);
    Console.WriteLine(res.TextBody);
}

Additional context
Add any other context about the problem here.

The expected behavior is really going to depend on the particular message you are parsing.

That said, it looks like you took it upon yourself to debug this issue and find the cause (thanks!).

It definitely looks like that logic could use some improvement.

Sorry @jstedfast I was away for a while and I'm only able to test latest release(4.4.0), I still have problems getting MimeMessage.BodyParts correctly, it looks like text/plain and text/html body parts are skipped when we set depth to 1.

Here is what I get in debugger with this change:
new

Here is what I was getting before:
old

Set MaxMimeDepth = 2

When MaxMimeDepth = 1, the multipart/alternative is treated like a MimePart instead of a Multipart.