Hyperlink detection incorrectly matches trailing characters

[Torbjörn Lönnemark]

Currently, enclosed URLs, like the following examples:

* «http://example.com/»
* <http://example.com/>
* (http://example.com/)
* "http://example.com/"

will match the trailing enclosing character, causing the Open link action to open an incorrect URL.

sakura will also match punctuation (e.g, '.', ',') at the end of URLs, which commonly occurs in written sentences. Matching these is almost never what you would want nor expect.

gnome-terminal handles both of these classes correctly.

For reference: https://bugzilla.gnome.org/show_bug.cgi?id=756038.

[Torbjörn Lönnemark (tobbez)]

It would be possible to achieve this by copying gnome-terminal's src/terminal-regex.h, including that in sakura.c and replacing the current HTTP_REGEXP with REGEX_URL_AS_IS.

[Markus Kurtz (mgkurtz)]

This bug seems not to be fixed. The http regex still matches the URLs given in the bug description and the mail regex matches '<<email address hidden>'. Simple regexes that should work in most cases would be

#define HTTP_REGEXP "(ftp|http)s?://[^ \t\n\b]+[^.,!? \t\n\b()<>{}«»„“”‚‘’\\[\\]\'\"]"
#define MAIL_REGEXP "[^ \t\n\b()<>{}«»„“”‚‘’\\[\\]\'\"][^ \t\n\b]*@([^ \t\n\b()<>{}«»„“”‚‘’\\[\\]\'\"]+\\.)+([a-zA-Z]{2,})"

There are more complex regexes around for this job, like the different versions at
https://daringfireball.net/2010/07/improved_regex_for_matching_urls

[David Gómez (dabisu)]

Thanks Markus, I've updated the regular expressions.