Sometimes I copy e-mail address with 'anti-spam' features, such as myname(at)myisp.com or myname myisp.com where an image of a @ used to be in the middle instead of email@example.com And it's annoying for me to correct them. How about having that done automagically?
The whole reason people use those anti-spam munged addresses is that they're supposed to be difficult to parse. xref bug 331116
See also Thunderbird bug 317320.
*** Bug 317320 has been marked as a duplicate of this bug. ***
At the dupe, Hans-Andreas Engel wrote: > Interestingly, the fact that the suggested transformation can be automated > means that spammers can perform it automatically as well, proving that this > attempt to avoid spam does not work well in the first place. > > But because not everyone reads this argument, the suggested feature remains > useful. Providing support only for the simple substititions of the " dot " => "." and " at " => "@" cases would certainly be possible; BUT, once any recognition was in place, then every variation -- parentheses, brackets, text-to-be-excluded -- would be another 'feature' that people would ask for, until the munged-address recognizer required an AI to implement. Further, providing a feature that allows, even encourages, people to continue with this pointless measure is counterproductive.
(In reply to comment #4) I can see where you're coming from... still, I find this useful enough to have at a rudimentary level. We can limit it, say, to the application of a few regular expressions (s/foo/bar) to the pasted address.
*** Bug 349981 has been marked as a duplicate of this bug. ***
> Providing support only for the simple substititions of the > " dot " => "." > and > " at " => "@" > cases would certainly be possible; BUT, once any recognition was in place, then > every variation -- parentheses, brackets, text-to-be-excluded -- would be > another 'feature' that people would ask for, until the munged-address > recognizer required an AI to implement. Most of the variations could be fixed by applying: s/([^A-Za-z0-9_-]+)dot([^A-Za-z0-9_-]+)/./ s/([^A-Za-z0-9_-]+)at([^A-Za-z0-9_-]+)/@/ (Am I missing any other legal characters in the substitution?)
Maybe s/([^A-Za-z0-9_-]+) [\[\(]?at[\]\)]? ([^A-Za-z0-9_-]+)/./ plus the application of the first regexp might need to be conditioned on the presence of an @ somehwere (or the application of the second).
Actually, I should add to my comment from 2007 that the pattern for an email address is more complex than that. You can have all sorts of strange characters, plus the rule for a first character is different from the following ones etc.
This might be useful as an extension. I don't think it deserves energy in core unless tested in an extension. It also seems to undermine the very purpose of munged addresses, and encourage their use which is also arguable. Perhaps it should be wontfix.