On Sun, Dec 25, 2011 at 19:18, Daniel Stenberg <[email protected]> wrote: > On Sat, 24 Dec 2011, Alex Vinnik wrote: > >> I am having a problem using libcurl in my Ruby app (Curb gem binds >> directly to libcurl). Specifically libcurl can't follow a weird redirect >> served by a web server. Server Location header "Location: >> http://www.officedepot.com;jsessionid=0000EXsMRFMF5kwJo26qgOif31d:13ddq0tfm"; >> doesn't look RFC complaint to me. For whatever reason jsessionid gets added >> to the end of new location. > > > Wow. That's a broken URL that I've not seen used before. [..] > > Since this is a fairly big site and at least one of the major browsers > support this format, I think we should consider supporting it. Even though > it would be under protest. > > Any other opinions?
My gut feeling is that I would rather campaign against Firefox supporting that format.. I've tried other browsers, and (in addition to Chrome) none so far will accept it. If Firefox keeps accepting any and all formats where no thought has been give to the implications then we'll just end up with a useless ad-hoc standard for URLs which just gets harder and harder to implement in an efficient way. (And gmail doesn't interpret the full string as an URL either - URL-parsing gets more difficult with such ad-hoc schemes too). -Tor ------------------------------------------------------------------- List admin: http://cool.haxx.se/list/listinfo/curl-library Etiquette: http://curl.haxx.se/mail/etiquette.html
