Hello, World

[see pwlin's comment]

Test bolduserscripts.org

Just a note, now that I have had Jesse unlist this test script... I will "ignore" any user input regarding it's assumed effectiveness.

LOL voted no on my own test script and it incremented the yes fields.

A terrible attempt in how 'not' to write a userscript, while in the same time fails to adhere to the most basic rule of using a unique @namespace. Please remove it before more innocent souls are affected by this wrong example.

Fails catastrophically at what it attempts (show the proper ways of user script crafting), for suggesting a namespace of http://localhost which fails the basic idea of @namespace: to be some identifier (any identifier!) that is _unique to you, as a user script author_, so Greasemonkey won't replace an already installed script named something one of your userscripts are named, on installing it. If everybody used "http://localhost" as their @namespace, no two user scripts could share a @name and be installed at the same time. Avoid!

Piontless

Vague, also provides inaccurate information.

Fails to describe metadata accurately. It doesn't describe what kinds of values @include can accept, or how the * wildcard works (e.x. greediness). Fails to mention which metadata fields are necessary and which are non-standard (and thus completely ignored by Greasemonkey).

Fails to mention issues like the effect of XPCNativeWrappers on the usage of expando properties, etc.. Fails to describe GM APIs. Fails to describe how to access variables on the page (unsafeWindow), and fails to explain the danger of relying on them (unsafeWindow).

Erroneous comment about anonymous functions and Greasemonkey version 0.8. Version 0.8 of Greasemonkey (which I'm actually using and coding for right now) does not throw all scripts in the same JavaScript namespace, so you don't need to wrap your code in an anonymous function. I've actually gone through a bit of GM's source code; what it does is, it places each script in an anonymous function for you, and then runs the code in its own sandbox (a new global scope -- think of a Window object with no default methods, properties, etc.), completely and totally separate from all other scripts, be those scripts userscripts or scripts on the page.

Erroneous demonstration of the @namespace metadata field. A very, very bad recommendation for its value is made: http://localhost. The field is used as a unique identifier for authors. Let's say two people named "John Smith" create separate scripts named "YouTube Downloader". Now, let's say you install one of the "YouTube Downloader"s, and then try to install the second. Greasemonkey now has a dilemma: it needs to figure out if the second script is a different version of the first script, or if it is a different script altogether. Now, the name and author are the same, so it can't check that; therefore, Greasemonkey is coded to use the @namespace identifier. So the identifier is non-semantic (in that you can specify whatever the hell you want for a value -- a URL, your school cafeteria PIN number, an e-mail address, whatever), but it still has a purpose (uniquely identifying the script's author), so it's foolish to just toss http://localhost into the field and misleading to recommend doing so -- if you're gonna use a non-unique identifier for a field that's supposed to have a unique value, you might as well not even specify anything for the field at all.

Testing: can we use HTML?

codeem

Marti wrote:

Newlines test