What's the current thinking on the long-URL problem? I posted about it ages ago, and ita, did I email you some code?
My thoughts on possible ways to handle long URLs in posts were, if I recall correctly:
[too-long URL removed]
- change them to
http://domainname/[rest of too-long URL removed]
- change them to
http://domainname/[...]/porn.htm
that last one being really "retain some of the URL if we can to provide a bit of context", but not really thought out.
I like the third idea. But, then, I don't have to code it.
I don't think the last one is a safe solution, since querystrings can make that still a very long adjusted URL.
I just had my first weird bug-that-can-only-be-discovered-through-random-user-behavior. I used the quick-edit spoiler font, and then started the next line with an arrow, using a caret to create the point of the arrow. All the text following the caret got eaten.
I don't think the last one is a safe solution, since querystrings can make that still a very long adjusted URL.
I think the formula in my head is "remove all directory information after the domain name, and remove all query string stuff after the name of the document, if there is any"
so
www.somewhere.com/blah/blah/blah/blah/blah/porn.htm
becomes
www.somewhere.com/[...]/porn.htm
and
www.somewhere.com/blah/blah/login.php?x=123456&y=123456&z=123456
becomes
www.somewhere.com/[...]/login.php
or something like that.
It's the query strings that are normally the problem, not the dirs, right?
Starting with John H.'s post 501 in Buffistatech, everything is aligned with the bullets he made. Which means the page alignment is off.
I like the third option, because I like if possible to know what file type is being linked to, so that I don't click on a 10MB .exe file by mistake. (Not that any Buffista would post such a thing without warning, but we have to plan for contingencies yada yada yada.)
John, as long as you regexp it, it's all good.
I think i broke the Firefly thread