Daniel Mayer wrote:
Thank you Brion! I guess it was my mistake to make the
original request while
you were on vacation...
One minor thing though; If it wouldn't be too much
trouble, then it would be great if the "Retrieved from" url text was an
active link.
http://www.wikipedia.org/w/wiki.phtml?title=Foo is also a bit
messy;
http://www.wikipedia.org/wiki/Foo would be better (this isn't that
important though).
Yeah, it kinda sucks. Right now what it does is it takes the exact URL
you gave it, in this case
http://www.wikipedia.org/w/wiki.phtml?title=Foo&printable=yes=yes, and
removes the &printable=yes.
I've checked a fix into unstable to use the canonical URl for the page,
but hypothetically it may not work in all circumstances, like if you
want a printable version of an old page revision; it wouldn't include
the version number, and we might want that.
Running on
test.wikipedia.org.
A related problem, which I dunno how to usefully fix, is that URLs for
pages in non-latin charsets are really REALLY ugly and long. Between
UTF-8 and ASCII-based URL encoding, one character baloons to six or even
nine chars of incomprehensible digits and percent signs. It doesn't much
matter when it's just between your browser and your bookmark file, but
cut-n-paste is painful, and printing is a real problem.
A moderately long title may not even fit within the width of a printed
page, like:
http://ru.wikipedia.org/wiki/2003_%D0%9A%D1%80%D1%83%D0%BF%D0%BD%D0%B0%D1%8…
We could try to give the URLs without url-encoding (where possible), and
hope that the users are able to type all characters and they'll all be
sent in an encoding the wiki can decypher, but I don't know how reliable
this will be.
-- brion vibber (brion @
pobox.com)