I seem to recall that this has been on Eli's mind more some time and that he has something rearing to go.<div><br></div><div>I think that is probably a good idea and that this breakage here is a bit too dangerous.</div> <div><br></div><div>I'll revert the commit and put a new HTTP library on my list to code and discuss with Eli.</div><div><br></div><div>Jay<br><br><div class="gmail_quote">On Wed, Nov 23, 2011 at 11:08 AM, Robby Findler <span dir="ltr"><<a href="mailto:robby@eecs.northwestern.edu">robby@eecs.northwestern.edu</a>></span> wrote:<br> <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Oh, right. Pattern matching on url structs. If we want to keep that<br> working (which I think we do), then that ties our hands much more.<br> <br> I believe there are some other crufty things in net/url (having to do<br> with encodings?).<br> <br> Does it make sense to have a new library that does all of this right<br> from the start and then announce that we're going to stop supporting<br> net/url in a year or so, in favor of the new library?<br> <br> I see there is no net/http. so we could also do a better job with the<br> whole simple-script-for-downloading-a-file-over-http and perhaps other<br> things along those lines?<br> <font color="#888888"><br> Robby<br> </font><div><div></div><div class="h5"><br> On Wed, Nov 23, 2011 at 12:05 PM, Jay McCarthy <<a href="mailto:jay@racket-lang.org">jay@racket-lang.org</a>> wrote:<br> > I thought about that too since there are few instances where people pattern<br> > match on the URL struct. What would be a good name for the new field...<br> > url-maybe-query?<br> ><br> > On Wed, Nov 23, 2011 at 11:02 AM, Robby Findler<br> > <<a href="mailto:robby@eecs.northwestern.edu">robby@eecs.northwestern.edu</a>> wrote:<br> >><br> >> I don't think we want to change how the current url struct selectors<br> >> work when applied to url structs.<br> >><br> >> You could probably get away with changing the url struct if you could<br> >> provide functions that act the way the old selectors used to work --<br> >> does that help?<br> >><br> >> Robby<br> >><br> >> On Wed, Nov 23, 2011 at 12:00 PM, <<a href="mailto:jay@racket-lang.org">jay@racket-lang.org</a>> wrote:<br> >> > jay has updated `master' from 6a99c93ebb to 9d8d36e568.<br> >> > <a href="http://git.racket-lang.org/plt/6a99c93ebb..9d8d36e568" target="_blank">http://git.racket-lang.org/plt/6a99c93ebb..9d8d36e568</a><br> >> ><br> >><br> >> ><br> >> > 7f9818b Jay McCarthy <<a href="mailto:jay@racket-lang.org">jay@racket-lang.org</a>> 2011-11-23 10:35<br> >> > :<br> >> > | This fixes 10497 and potentially breaks programs that assume the query<br> >> > of a URL is always a list. I have fixed uses in the Web Server, which I<br> >> > expect is the major thing affected, but much more could be. Therefore I am<br> >> > skeptical this is a good idea just for the representation of ?. So, I'd like<br> >> > other people to review the change and let me know if they think I should<br> >> > revert it.<br> >> > :<br> ><br> ><br> </div></div></blockquote></div><br></div>