W3C home > Mailing lists > Public > www-tag@w3.org > July 2003

Re: Invalid JAR URIs

From: Bill de hÓra <dehora@eircom.net>
Date: Thu, 03 Jul 2003 22:21:15 +0100
Message-ID: <3F049E4B.6000004@eircom.net>
To: Chris Lilley <chris@w3.org>
CC: www-tag@w3.org

Chris Lilley wrote:

> Sure, but if its not a URI then its useless inside the JDK as well.

Demonstrably not true. These things are useful, they're just not URL/Is.


> Details welcome.

Well, there was no URI class until 1.4.

Historically URL and the support classes particualry 
HttpURLConnection have had (imo) embarassing bugs - the word 
'notorious' comes to mind. This isn't relevant to the jar: 'scheme', 
but as you asked, some of the problems have been or are:

  url/i equivalence
  cookie handling
  parsing urls
  redirects
  webdav support awkward
  encodings
  buffered streams and blocked reads
  keepalive
  sometimes stupifyingly slow on windows compared to a socket

that's just the stuff that's bitten me. the list of things that were 
fixed in 1.3 and 1.4 are much longer:

  http://java.sun.com/j2se/1.3/fixedbugs/1.3.0/classes_net.html
  http://java.sun.com/j2se/1.4/fixedbugs/fixedbugs-all.html

But the main point is this - even if this was real URI, the JAR 
support code simply does not require a jar:http: prefix - http: 
would do just as well (or even jar: )

Bill de hÓra
Received on Thursday, 3 July 2003 17:25:17 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:59 UTC