[Libreoffice-bugs] [Bug 130501] URL in unicode interpretate as file link (PDF export)

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=130501

Stephan Bergmann <sbergman at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sbergman at redhat.com

--- Comment #3 from Stephan Bergmann <sbergman at redhat.com> ---
There appears to be an issue with internationalized domain name top level
domains (IDN TLDs) somewhere in the code:

The IDN URL <http://транссеть.рф> (with a IDN TLD) is exported as a URL with a
last segment of "http:%2F%2Fxn--80akxkhacg4g.xn--p1a%D1%84" (i.e., some code
apparently decoded IDN label "транссеть" to "xn--80akxkhacg4g" and "рф" to
"xn--p1a%D1%84", though the latter should have been "xn-p1ai"?!?, and then
URL-encoded "//" as "%2F%2F" as it apparently decided for some reason that all
of the input should become a single segment in the resulting URL).

On the other hand, testing with current master on Linux at least,
<http://транссеть.ru> (with a "plain" TLD) is exported as
<http://xn--80akxkhacg4g.ru/> (i.e., with the IDN label "транссеть" decoded to
"xn--80akxkhacg4g", which should be fine).

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20200207/3e2f0354/attachment.htm>