Comments on: Percent Escapes Gotcha http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/ my weblog. Wed, 17 Mar 2010 19:41:03 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Vincent Gable http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/comment-page-1/#comment-621 Vincent Gable Sun, 12 Apr 2009 00:26:57 +0000 http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/#comment-621 "What’s wrong with just using the normal encode/decode functions and being careful to use each once? " I don't like being careful if I don't have to :-). But seriously, if I go down that route I will probably make a <code>PercentEscapedNSString</code> object, that just holds an <code>NSString</code> and on creation percent-escapes whatever string it's given. That way <strong>I can let the type-system make sure I'm doing things right</strong>. “What’s wrong with just using the normal encode/decode functions and being careful to use each once? ”

I don’t like being careful if I don’t have to :-). But seriously, if I go down that route I will probably make a PercentEscapedNSString object, that just holds an NSString and on creation percent-escapes whatever string it’s given. That way I can let the type-system make sure I’m doing things right.

]]> By: Vincent Gable http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/comment-page-1/#comment-620 Vincent Gable Sun, 12 Apr 2009 00:20:43 +0000 http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/#comment-620 Good catch Jason, if someone types in something that has a valid percent escape, eg "cat%20dog", then I will <em>incorrectly</em> un-encode it it into "cat dog" by <code> stringByReplacingPercentEscapesOnce</code> and send a URL with "cat%20dog" <em>not</em> "cat%2520dog". My <a href="http://vgable.com/blog/2009/04/10/pre-announcing-prometheus/" rel="nofollow">application</a> takes text typed into a search-box, and builds a search-URL from it. It also takes the URL of a wikipedia page, extracts the page name from it, un-mangles it, and puts it in the search box (which conceptually acts like the URL/search bar in <a href="http://www.google.com/chrome" rel="nofollow">Chrome</a>.) Wikipedia has a nice property that if your search-string matches a page name, you go right to that page name. So I'm not sure if this is anything but an academic problem, unless there are page names that have percent-escapes in their real name. Still, it looks like I probably will not be reusing these functions anytime soon. Good catch Jason, if someone types in something that has a valid percent escape, eg “cat%20dog”, then I will incorrectly un-encode it it into “cat dog” by stringByReplacingPercentEscapesOnce and send a URL with “cat%20dog” not “cat%2520dog”.

My application takes text typed into a search-box, and builds a search-URL from it. It also takes the URL of a wikipedia page, extracts the page name from it, un-mangles it, and puts it in the search box (which conceptually acts like the URL/search bar in Chrome.) Wikipedia has a nice property that if your search-string matches a page name, you go right to that page name.

So I’m not sure if this is anything but an academic problem, unless there are page names that have percent-escapes in their real name.

Still, it looks like I probably will not be reusing these functions anytime soon.

]]> By: Jason Petersen http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/comment-page-1/#comment-619 Jason Petersen Sat, 11 Apr 2009 22:34:45 +0000 http://vgable.com/blog/2009/04/10/percent-escapes-gotcha/#comment-619 My inability to read Obj-C syntax prevents me from understanding your method. How exactly can an idempotent version of an encoder without making some heuristic assumption about what 'encoded' looks like? I mean if someone has a fairly simple string it's fairly easy, but what if I WANT to encode this string: "100% escaped once: 100%25, escaped twice: 100%2525" to be properly sendable in a URL. What's wrong with just using the normal encode/decode functions and being careful to use each once? Outgoing, encoding needs to be the last thing you do. Incoming, decoding is the first. My inability to read Obj-C syntax prevents me from understanding your method. How exactly can an idempotent version of an encoder without making some heuristic assumption about what ‘encoded’ looks like? I mean if someone has a fairly simple string it’s fairly easy, but what if I WANT to encode this string:

“100% escaped once: 100%25, escaped twice: 100%2525″

to be properly sendable in a URL. What’s wrong with just using the normal encode/decode functions and being careful to use each once? Outgoing, encoding needs to be the last thing you do. Incoming, decoding is the first.

]]>