We had this exact problem and this fixed it for us. It would be great to have this merged and a new version released.

@tenderlove the last commit in this gem was yours, can you check this PR please?

Can we get these merged in?

We also have exactly this case patching now. Can anyone please merge this in?

@mfbmina Sorry for lack of attention here, but thanks for opening this PR!

I noticed that the added test case seems to pass with or without the change, probably because the two calls share an implementation.

If you're still game for it, could you update the test case to be based on a string literal, similar to other nearby tests? I'll do what I can to facilitate incorporating the patch into a new release. 🙇

@mattyoho is the only thing blocking this merge the test case? If I fixed it could this be merged?

@mattyoho @MelizzaP tests are fixed now! Sorry for taking so long 😄

@mfbmina @MelizzaP Thanks for taking a look at this PR again!

Appreciate the updated test, @mfbmina. 👍

Unfortunately, coming back to this pull request, I need to ask for some further changes to how the patch is structured before merging, because it works differently than the existing code for detecting a quote header does.

Here in this PR we're removing the Google Mail quote header from the incoming text on the "first pass" that performs some line normalization prior to the "second pass" of line-by-line scanning. (I use quotes here because the code isn't explicitly structured as a multi-pass implementation, though de facto it is.)

But the existing quote header detection leaves the header text in place so that it's subsequently included in a quote fragment on that second pass.

Adjacent to the addition here, the existing detection just normalizes the header text to remove any newlines that may have been introduced by the line-wrapping conventions for email messages:

It then later uses quote_header? to decided whether to also include a line not preceded by > with a quote fragment being built:

Where the definition of quote_header? uses a simplified version of the regular expression used during the normalization pass to remove inserted newlines:

So what I would ask is for two changes:

1. Rather than remove the Gmail quote header text, normalize it to remove inserted newlines, as is being done for the existing header detection.
2. Update the quote_header? implementation to also detect the Gmail format.

This will have the benefits of not stripping the header text and also associating the newline that follows it with the quote block.

With those changes I think we can merge and then cut a pre-release version of the gem that includes the patch. Happy to discuss things further if needed.

Thank you for contributing!

@mattyoho I applied the changes that you required! Hope that solves all issues 👍

thanks @mfbmina and @mattyoho 😄! This fixes an issue my team is having, any estimated timeline on a merge and new release?

@mattyoho do you had a chance to look?

@mfbmina is this ok to merge?

@MelizzaP I think so... Tests are passing and I did the requested changes.

whoops, sorry, I tagged the wrong person. @mattyoho is this ok to merge?