> As you suspected, the problem went away when I manually set > all the | to %7C. Perfect fix! Thanks for that info.
No worries. > Caused by: java.net.URISyntaxException: Illegal character in > path at index 99: > I had already used the error message to zero in on the > separator bar character. But you didn't go that one step further, and decide that it was an illegal character. ;) (I'm just kidding. As you say, if you don't have any experience with HTML/HTTP there is little reason for you to be aware of this, but you can see how the exception gives me more information than your original post. Laurens (the developer) can get even more information than I can out of the exception. It will give him a reasonable idea as to what the code was doing when you hit the error.) > Here come the inevitable picky questions: When Jpluck "add > site" automatically pastes in the URI that is in your > clipboard, why not correct the character mappings to a > Jpluck compatible URI? Now that's a question that only Laurens can answer. Or you could download the source to JPluck, and try to figure out where the extra code would go. I'm sure he would be more than happy to accept a patch enabling this functionality. (Although, perhaps the patch should do the conversion when the user hits the "Okay" button, to avoid confusion.) > If Jpluck is going to be more restrictive than Internet > Explorer (which takes this http address just fine), then > that might be a valid change to make. Personally, I very much hope that JPluck continues to be more restrictive than IE. IE contains a ton of code to work around people who write invalid HTML, and I would hate to see JPluck bloat in this way. It's already much more forgiving of errors than the Python parser, from what I remember. Being unrestrictive comes with costs. Costs that most users don't seem to recognize or understand. Everyone complains about software getting bigger and slower, but no-one seems to realize that it's partially due to the developers trying to compensate for invalid input from the very users who are complaining. > Also, when Jpluck "add site" fails on this URI, it does so > silently. It simply doesn't add the document to the list. > That's not helpful. Speaking as a developer, specifying the behaviour you would like to see is much more useful than just saying that the current behaviour is no good. Further, supplying a patch to the current code that implements your new beahviour is light-years better than describing it. We realise that not everyone is a developer, and so don't hope for the second option, but everyone should be capable of the first. Speaking as a user, I completely agree. Ideally, from my perspective, the offending character in the url should be flagged in red, and the error message printed in red above the field. But that might be a little tricky. Later, Blake. _______________________________________________ plucker-list mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-list
