WSDL Support in Mozilla 1.4 Final

Tuesday May 20th, 2003

Doron Rosenberg writes: "Netscape has been working on improving the web services support in Mozilla, and as of 1.4 final, Gecko will support WSDL 1.1 proxying. View the translating example here and find out more about the effort at the project page.

"Please note that you need trunk builds of Seamonkey or Firebird (Firebird 0.6 is OK) or else the browser will crash. We have a 'Web Services' component in Bugzilla under 'Browser' to file bugs against.

"Try figuring out what the following Chinese phrase means: Mozilla 伟大!"

#1 it means...

by ZoSo

Tuesday May 20th, 2003 12:49 AM

..."Mozilla is great".

#2 Re: it means...

by ZoSo

Tuesday May 20th, 2003 1:15 AM

Oops...misread, didn't mean to give the game away. Is it possible for the admins to delete or edit my post?

#3 Crash

by Mazon

Tuesday May 20th, 2003 1:18 AM

So I can crash Mozilla with a simple script?

#7 Re: Crash

by mlefevre

Tuesday May 20th, 2003 2:13 AM

Well you can if you've got an older version. But not any more with this particular example, however :)

It's not hard to crash or freeze mozilla (or IE, for that matter) from scripts - there are a bunch of other such bugs filed in bugzilla waiting to be fixed. A particularly tricky one is infinite loops in scripts - detecting those is a well known problem in computing.

#8 Re: Crash

by Mazon

Tuesday May 20th, 2003 2:35 AM

yeah, it is a bit tricky to detect a rogue script - but using a valid script should never crash a browser!

#9 no it shouldn't

by niner

Tuesday May 20th, 2003 3:09 AM

and this is the reason many people are working constantly on findig and fixing such bugs. Of course in a project consisting of millions of source code lines this takes some time and will probably never be finished. But it's getting better and better and in most cases the user will hopefully never stumble across such a bug.

#12 It usually doesn't

by Mark_Watkins

Tuesday May 20th, 2003 5:28 AM

I've been using various Gecko browsers intensively for years now and don't remember having a Javascript crash, certainly not in recent history. I'd say even if you're doing bleeding-edge javascript (would love to myself, but need someone to pay me to do it ;-) ) you'd have to try pretty hard to get a crash.

#13 Try this:

by mw22

Tuesday May 20th, 2003 8:17 AM

<html> <fieldset style="position:fixed;"> <legend>Crash</legend> </fieldset> </html>

So little code, but so much fun...

#21 Re: Try this:

by Mazon

Tuesday May 20th, 2003 3:40 PM

Oh man, thats bad. And I laughed at the IE bug that made it crash with some garbled html tags :/

#26 Valid scripts and infinite loops

by hodeleri

Tuesday May 20th, 2003 5:41 PM

There is no general process to determine if any given function with any given input will terminate or not. If you can prove to the contrary, great fame will be deservedly yours.

Here's another that is perfectly valid, but crashes: data:text/html,<script>document.write(document.location);</script>

If a language engine can't even tell if its inputs will terminate or not, how can it tell if the code is executing is 'rogue' or not?

#27 Re: Valid scripts and infinite loops

by bzbarsky

Wednesday May 21st, 2003 12:44 AM

> If you can prove to the contrary, great fame will be deservedly yours.

Especially since the fact that there is no general process _has_ been proven. Because once you prove both things, you should be able to trivially derive that 1 == 0 and from then on, the world is your oyster.

#34 Re: Valid scripts and infinite loops

by chithanh

Tuesday June 3rd, 2003 7:53 AM

To be exact, it has been proven that no Turing machine can solve this problem, but IIRC it might still be possible to invent your own (more powerful) computational formalism which makes it possible to find out whether a program terminates or not.

Unfortunately, even then you won't be able to implement this on your computer.

#14 Only crashes on older trunk builds

by doron

Tuesday May 20th, 2003 10:24 AM

It crashes if you have anything after 1.3 final went out, the trunk. No final release will crash, as the code wasn't turned on yet.

#4 Unicode characters?

by gregk

Tuesday May 20th, 2003 1:37 AM

If this page includes Unicode entities, shouldn't it be UTF-8 rather than ISO-8859-1?

#5 Re: Unicode characters?

by james

Tuesday May 20th, 2003 1:54 AM

The encoding of the text _is_ ISO-8859-1. If they set the encoding to UTF-8, there wouldn't have been any need to use character references at all. The main reason for using character references is to include characters that can't be represented in the document's charset.

#6 Unicode characters?

by gregk

Tuesday May 20th, 2003 2:07 AM

Does ISO-8859-1 include characters numbered decimal 20255 and 22823?

#10 No, but Unicode does

by jsoderba

Tuesday May 20th, 2003 3:32 AM

From the HTML 4.01 Recommendation, chapter 5

"HTML uses the much more complete character set called the Universal Character Set (UCS), defined in [ISO10646]."

"Numeric character references specify the code position of a character in the document character set."

"The syntax "&#D;", where D is a decimal number, refers to the ISO 10646 decimal character number D."

#11 Re: Unicode characters?

by james

Tuesday May 20th, 2003 5:07 AM

No, which is why a character reference was used to encode those characters in the story.

#16 Re: Unicode characters?

by tny

Tuesday May 20th, 2003 12:41 PM

There are two issues here. You have to distinguish between the Unicode standard and the Unicode Transformation Format encodings. You can use Unicode in any encoding with entities, even if that encoding does not include codepoints for those Unicode characters, provided that your application supports Unicode entities. You do not need to use entities for Unicode characters in the UTFs because all Unicode characters are represented by the UTFs.

An example: if you want to use one of the Han characters from Unicode on an ISO-8859-1 page, you need to reference it with an entity because the character is not part of that encoding - it has no assigned code point in that encoding (which, after all, only has 256 code points). But the same character can be be referenced in UTF-8 *without using an entity* because it has an assigned code point in UTF-8.

#23 Re: Unicode characters?

by choess

Tuesday May 20th, 2003 4:14 PM

No. HTML provides access to Unicode via numeric character entities, so that glyphs from the Unicode character set can always be displayed. The misleadingly-named "charset" actually applies to the character *encoding*; that is, the algorithm used to transform the binary (or octal, or...) representation of the document into human-language characters. Since the numeric character entities are written in the document in ASCII characters (a subset of ISO-8859-1), there's no need to change the charset.

#15 Well, here is how it went for me....

by stephenc

Tuesday May 20th, 2003 12:39 PM

I have Mozilla Firebird 0.6 on MS Windows 2000.

I came to this website and a dialog box popped up asking me to install Chinese. I don't speak Chinese and thought this was a glich, so I clicked on the cancel button. Waited. Clicked cancel button again. Waited. Clicked "X" box in upper right of dialog box to close it. Saw this topic. I don't think the Cancel button works on the dialog box.

And the Chinese phrase "Mozilla ??!" probably means "Mozilla, huh? what? Oh, its not made by Microsoft!" (um, see you just forgot to stick in the rest of the words....)

#17 Re: Well, here is how it went for me....

by tny

Tuesday May 20th, 2003 12:43 PM

You didn't have Chinese display installed, and the page includes Chinese characters.

#18 Re: Well, here is how it went for me....

by Ascaris

Tuesday May 20th, 2003 1:08 PM

Given that most of us that read this English language web site do not read Chinese, it is pretty annoying to have Chinese characters on the page, thus making every one of us that does not have Chinese installed (and does not need it, because it looks like ??! to us anyway) answer that dialog that asks us to install Chinese. If we want to see an example of the translating, we can follow a link... that should not be on the entry page.

#19 You'll want to vote for...

by Sander

Tuesday May 20th, 2003 1:54 PM

...bug 178183 - - and more importantly bug 86525 - - to get rid off the stupid dialog once and for all. I don't have chinese (or japanese, or korean) installed, I didn't want it installed _last time_ I visited a page using such characters, and I don't want to install it _next time_ either!

#20 Woo!

by Sander

Tuesday May 20th, 2003 3:35 PM

...and Asa just set blocking1.4+ to that first bug. I'd have preferred the second, but still... *does the happy dance*

#28 excellent!

by niner

Wednesday May 21st, 2003 2:35 AM

Just that such a litte bug with a very easy workaround (just click on x) may block the coming release shows how far Mozilla has come. This is quality work and nothing else :)

#31 Re: Woo!

by ecarlson

Friday May 23rd, 2003 10:30 PM

That's great. That darn dialog box, with the non-functioning cancel button has been annoying me for a very long time (so much so that I finally installed the simplified Chinese code set.) And why does it say "cancel"? Shouldn't it say "Close" or something like that, since you're not canceling anything, just closing a message box.

At least now I can actually see the message in the box. For a while the box was sized so small that all I could see was the title bar buttons, so I didn't know what the message was about.

- Eric,

#33 Duh, I already knew that

by stephenc

Tuesday May 27th, 2003 6:32 PM

I can read just fine thank you. The point of my post is that the Cancel button didn't work. I didn't want to download a Chinese font because I'm not interested in viewing Chinese. Since I don't read, write, or speak Chinese a bunch of puncuation marks are just as meaningful. The only reason I came to this topic in the first place was to find out what the acronym "WSDL" meant.

#29 Re: Well, here is how it went for me....

by GAThrawn

Wednesday May 21st, 2003 5:45 AM

But if you read what the box says and follow its directions it goes away forever AND lets you see those pretty Chinese characters.

#32 You made some assumptions

by stephenc

Tuesday May 27th, 2003 6:26 PM

You assumed that I would want to see Chinese characters. I don't becuase they have no meaning to me. I also don't want to download fontpacks for something I'm not interested in. Mozilla is unstable enough, anyway.

#22 This is incredibly cool!

by dpol

Tuesday May 20th, 2003 3:46 PM

Wow, I have waited for something like this! While I'm not convinced that HTML is appropriate for building applications (after all, it's supposed to be a way to exchange informations with simple computer-understandable semantics), I'm still quite impressed.

I remember back in the day, when I was writing a web interface to a standard application. I wanted to retrieve data from the server using JavaScript, without forcing a page reload, and then update the DOM accordingly. I tried to do this by dynamically creating a new document and loading data into it. Unfortunately, browsers refused to actually load the data, as the document was invisible.

I wish I had access to this technology back then.

#24 Fonts?

by WillyWonka

Tuesday May 20th, 2003 4:51 PM

How come the dialog box doesn't offer to go to a site that has a downloadable font? (Or does it? I haven't seen it since this morning)

#30 Re: Fonts?

by GAThrawn

Thursday May 22nd, 2003 2:25 PM

As I mentioned above the dialog box gives you straightforward, simple instructions on how to install Simplified Chinese font support to your OS without any need for downloads as long as you have the original Windows install CD or I386 install files somewhere.

(Well it does on Windows 2000 and XP anyway, couldn't promise for any others)

#25 congrats

by morg

Tuesday May 20th, 2003 5:05 PM

Just a quick word of congratulations to the developers. You guys are the best. You're keeping Mozilla at the leading edge. You're doing a great job, and we humble users really, really appreciate your hard work.

Thank you, again.