github / github/email_reply_parser
Signature regular expression
- Dominant language
- Ruby
- Stars
- 708
- Forks
- 156
- PR merge metrics
- No merged PRs in 30d
Description
I ran into some random situations where emails were not parsed correctly. Part of the email was chopped out, for example:
`Hi Test,\n\n-This is a test`
Will result in `-This is a test` fragment being hidden and considered as part of signature which it should not be. After digging more into the problem, we figured out that:
`(?m)(--\s*$|__\s*$|\w-$)|(^(\w+\s*){1,3} ym morf tneS$)`
The first matching group is testing against `\w-`, which is causing the problem. I can't think out of a better way to solve the problem other than take `\w-$` part out so it opens to any suggestion.
Contributor guide
Research direction
Start with the signature regular expression shown in the issue and reproduce the provided email example to inspect why the fragment is classified as a signature. Confirm the matching behavior after deciding how the \w-$ alternative should be handled, and verify that the sample fragment remains visible.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- ruby
- Domain
- backend
- Issue type
- Bug
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100