Body delimiters to truncate emails do not take uncommon whitespace into account
The function which truncate emails after a certain delimiter was found (i.e.
MailHandler#cleanup_body) currently does not take trailing whitespace into account that is not part of
\s. This e.g. includes non-breaking spaces which are common after e.g. signature delimiters (hyphen hyphen no-break-space).
This results in delimiters being not taken into account if there are these unicode whitespace characters at the end of the line.
The attached patch fixes this for Ruby versions >= 1.9. I have found no generally reliable way to fix this for Ruby 1.8 which thus retains the previous behavior. The reason this only works on Ruby >= 1.9 is that only this version includes a possibility to match unicode whitespace. Ruby 1.8 only has a limited list of ASCII characters (against the unicode spec apparently).
Given that I can't change the behavior on 1.8, this change is unfortunately rather difficult to test. I have also attached a patch with improved tests which work properly on 1.9 and 2.0 but breaks on 1.8 as it doesn't recognize the change in behavior. Given that, I'd rather prefer to pull the improvement of behavior without the test than to hold everything out...
#5 Updated by Marius BALTEANU over 4 years ago
Updated Holger Just first patch to apply cleanly.
#6 Updated by Go MAEDA over 4 years ago
- Target version changed from Candidate for next major release to 3.4.0
Thank you for fixing the patch. Now we can apply 0001-Allow-unicode-whitespace-after-mail-handler-body-del_r16545.patch and 0002-Add-test-for-trailing-unicode-whitespace-after-email.patch to the current trunk.
Setting target version to 3.4.0.