https://bugs.kde.org/show_bug.cgi?id=358049
--- Comment #5 from Erik Quaeghebeur <kdeb...@equaeghe.nospammail.net> --- (In reply to kdex from comment #4) Thanks for your reply. > I don't see a clear benefit from teaching people to misunderstand URLs. > Suggesting that they might not end in a dot will naturally break lots of > valid URLs, and the same goes for question marks (and potentially other > tokens that I haven't checked). Well, that is an URL entry issue. KMail could suggest people to encapsulate URLs it detects in mails about to be sent out using <...>. > Also, expecting that users will put punctuation symbols after their URLs to > end a sentence is a constructed heuristic; the majority of my sent and > received mails actually contain footnotes such as "[1]", which will be > completed with their URLs at the bottom of the mail, line by line. This > format is also very wide-spread and suffers from KMail's heuristic. Indeed this format is common. URLs ending with periods or question marks less so, and that is where it breaks. Nevertheless a valid point. > Next, observe that this is not just about usernames that might end in > punctuation. See [1] to agree that a dot at the end of a domain is, too, > valid syntax and should be parsed to reflect that. I would need to be convinced that this is a practical issue. I haven't come across mails with domains in this format. I have come across plenty of mails with periods as punctuation at the end of URLs, i.e., that won't resolve if the period is parsed as part of the URL. > In the case of usernames > or other dynamic parts in a URL, KMail *will* break websites (like [2], > which allows trailing dots in usernames and thus, in their profile page > URLs). Indeed; that is problematic. You've convinced me that this is a real issue. Nevertheless, always parsing periods as part of the URL will also break links. The KMail devs will have to decide what is most important. -- You are receiving this mail because: You are watching all bug changes.