Including other JS files and using "class" and custom styles

[Stefano POGLIANI]

Hi all
 I hope that, even if these are very basic questions, someone could give
me some hints.

   1. Does it work including an external Javascript script inside a GM
      script?
      For instance, if I want to include some "AJAX effects" and want to
      use one of the Open Source javascript libraries out there...
   2. In my GM script I am programmatically creating ALL the widgets
      that I need.
      Is there a way where I could use some "static HTML" to shortcut
      the development time? If yes, does someone have an example?
   3. one of the issues related to my poor knowledge of JS is that, when
      programmatically creating my widgets, I do not know how to set a
      custom CSS class (the equivalent of saying <div class="myclass">
      ..... how can I specify in Javascript this?)
   4. associated to the latter. If I create my own CSS classes, can I
      store their definition in a CSS style sheet file and "import" it
      in the GM script? If yes, how?

   5. Final one. Can I use XUL inside a GM script?

Thanks a lot for your help and patience
/Stefano

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Anthony Lieuallen]

On 8/4/2006 4:18 AM, Stefano POGLIANI wrote:
>   1. Does it work including an external Javascript script inside a GM
>      script?

The only straightforward way is to copy and paste the source in.
There's no "include" statement like in some other languages.

>   2. Is there a way where I could use some "static HTML" to shortcut
>      the development time? If yes, does someone have an example?

There's a number of ways.  For example:
http://www.arantius.com/article/dollar-e

>   3. one of the issues related to my poor knowledge of JS is that, when
>      programmatically creating my widgets, I do not know how to set a
>      custom CSS class (the equivalent of saying <div class="myclass">
>      ..... how can I specify in Javascript this?)

Since "class" is a reserved word in JavaScript, you have to use
something else.  In this case, className.

>   4. associated to the latter. If I create my own CSS classes, can I
>      store their definition in a CSS style sheet file and "import" it
>      in the GM script? If yes, how?

Yes.
http://diveintogreasemonkey.org/patterns/add-css.html

>   5. Final one. Can I use XUL inside a GM script?

That question is a bit too open ended for me to understand fully; I
can't be sure, because I don't know what you're really trying to do, but
I suspect the answer is no.
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Stefano POGLIANI]

Thanks Anthony!

Do you have an example of use of "className" in Javascript?

also, as I read the "Dive into GM" page, it seems that if the newly
defined "class" gets added an  "! important;" at the end, the definition
overrides the previous one.
Could this approach be used to "sort of" RESET any stylesheet attribute
that would apply to the elements I am defining in my page?
The issue here is to avoid that page-specific stylesheets affect the
display of my widgets in the page (so that on one page my DIV will paint
in one way and, on a different page, it will paint in  a different way).


As to the XUL question. I was wondering if, for instance, with GM I
could create a little XUL container which could, then, be embedded in
soething like the "All-in-One Sidebar" extension.

Thanks
/Stefano

Anthony Lieuallen wrote:

> On 8/4/2006 4:18 AM, Stefano POGLIANI wrote:
>>   1. Does it work including an external Javascript script inside a GM
>>      script?
>
> The only straightforward way is to copy and paste the source in.
> There's no "include" statement like in some other languages.
>
>>   2. Is there a way where I could use some "static HTML" to shortcut
>>      the development time? If yes, does someone have an example?
>
> There's a number of ways.  For example:
> http://www.arantius.com/article/dollar-e
>
>>   3. one of the issues related to my poor knowledge of JS is that, when
>>      programmatically creating my widgets, I do not know how to set a
>>      custom CSS class (the equivalent of saying <div class="myclass">
>>      ..... how can I specify in Javascript this?)
>
> Since "class" is a reserved word in JavaScript, you have to use
> something else.  In this case, className.
>
>>   4. associated to the latter. If I create my own CSS classes, can I
>>      store their definition in a CSS style sheet file and "import" it
>>      in the GM script? If yes, how?
>
> Yes.
> http://diveintogreasemonkey.org/patterns/add-css.html
>
>>   5. Final one. Can I use XUL inside a GM script?
>
> That question is a bit too open ended for me to understand fully; I
> can't be sure, because I don't know what you're really trying to do,
> but I suspect the answer is no.
> _______________________________________________
> Greasemonkey mailing list
> Greasemonkey@...
> http://mozdev.org/mailman/listinfo/greasemonkey

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Philip Friedman - Auto Europe]

Maybe not "straightforward" by Anthony's standards, but it does "include"
library.js in libraryTest.user.js as desired.

Here's the example, that works for the two files in the same directory on
the server. Note that libraryTest.user.js @includes itself, so the URL
construction won't make sense in the real world.

File libraryTest.user.js:

// ==UserScript==
// @name          LibraryTest
// @namespace     /phil
// @description   Library Test
// @include       http://*libraryTest.user.js
// @exclude      
// ==/UserScript==

GM_xmlhttpRequest( { method:'GET'
  , url:document.location.href.replace(/\/[^\/]*$/,'/library.js')
  , onload: onLoadLibrary
} );

function onLoadLibrary(rsp) {
  eval(rsp.responseText);
  libraryAlert();
}

File library.js:

function libraryAlert() {
  alert('libraryAlert');
  GM_log('libraryAlert');
}

--
Regards, Phil Friedman - Auto Europe - 207-842-2069
 

|-----Original Message-----
|From: greasemonkey-bounces@...
|[mailto:greasemonkey-bounces@...] On Behalf Of Anthony Lieuallen
|Sent: Friday, August 04, 2006 9:33 AM
|To: greasemonkey@...
|Subject: Re: [Greasemonkey] Including other JS files and using
|"class" and custom styles
|
|On 8/4/2006 4:18 AM, Stefano POGLIANI wrote:
|>   1. Does it work including an external Javascript script inside a GM
|>      script?
|
|The only straightforward way is to copy and paste the source in.
|There's no "include" statement like in some other languages.
|
|>   2. Is there a way where I could use some "static HTML" to shortcut
|>      the development time? If yes, does someone have an example?
|
|There's a number of ways.  For example:
|http://www.arantius.com/article/dollar-e
|
|>   3. one of the issues related to my poor knowledge of JS is
|that, when
|>      programmatically creating my widgets, I do not know how to set a
|>      custom CSS class (the equivalent of saying <div class="myclass">
|>      ..... how can I specify in Javascript this?)
|
|Since "class" is a reserved word in JavaScript, you have to
|use something else.  In this case, className.
|
|>   4. associated to the latter. If I create my own CSS classes, can I
|>      store their definition in a CSS style sheet file and "import" it
|>      in the GM script? If yes, how?
|
|Yes.
|http://diveintogreasemonkey.org/patterns/add-css.html
|
|>   5. Final one. Can I use XUL inside a GM script?
|
|That question is a bit too open ended for me to understand
|fully; I can't be sure, because I don't know what you're
|really trying to do, but I suspect the answer is no.
|_______________________________________________
|Greasemonkey mailing list
|Greasemonkey@...
|http://mozdev.org/mailman/listinfo/greasemonkey
|
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Anthony Lieuallen]

On 8/4/2006 9:42 AM, Stefano POGLIANI wrote:
> Do you have an example of use of "className" in Javascript?

Uh..
var div=document.createElement('div');
div.className='foo';

> also, as I read the "Dive into GM" page, it seems that if the newly
> defined "class" gets added an  "! important;" at the end, the definition
> overrides the previous one.

The "! important" clause relates to specificity:
http://www.w3.org/TR/CSS21/cascade.html#specificity
It makes that rule "more specific" than anything else.

> Could this approach be used to "sort of" RESET any stylesheet attribute
> that would apply to the elements I am defining in my page?

There's a GM api to do this .. I think it's only in CVS versions at this
point.  I can't remember its name, but someone will come up with it.

> As to the XUL question. I was wondering if, for instance, with GM I
> could create a little XUL container which could, then, be embedded in
> soething like the "All-in-One Sidebar" extension.

Can't say, for sure.
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[WillBo]

I can't remember what the status of this whole discussion is anymore.

Has anyone suggested another metadata statement, like "@library
userscript_library.user.js".
(since @include is already defined; the ".user.js" could be implied and
auto-appended if absent)
The desired file could be in the standard user script directory.
This would be fairly easy to do and support. (relatively, I think)
You could just "install" the library 'module', but not @include it to
anything, just to get
it into the directory for reference with the @library meta statement.

As for XUL, no, that is chrome-only, which you could do if you compiled
your GmScript
into an extension and then added the XUL stuff.

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Johan Sundström]

> I can't remember what the status of this whole discussion is anymore.
>
> Has anyone suggested another metadata statement, like "@library
> userscript_library.user.js".

It has been suggested, and there has even been floating around some
code to implement it. Unfortunately it still has not made it into a
released Greasemonkey version, for reasons I don't remember (they may
not have been aired publicly yet, for instance).

--
 / Johan Sundström, http://ecmanaut.blogspot.com/
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Aaron Boodman]

On 8/5/06, Johan Sundström <oyasumi@...> wrote:
> It has been suggested, and there has even been floating around some
> code to implement it. Unfortunately it still has not made it into a
> released Greasemonkey version, for reasons I don't remember (they may
> not have been aired publicly yet, for instance).

No particular reason, except that nobody has had time to do it. We
should not be adding random code (no offense) to GM just because a
patch exists.

Anyway, yes I know this is a very popular request. It's near the top
of the list of things to do.

- a
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Johan Sundström]

On 8/6/06, Aaron Boodman <zboogs@...> wrote:
> On 8/5/06, Johan Sundström <oyasumi@...> wrote:
> > It has been suggested, and there has even been floating around some
> > code to implement it. Unfortunately it still has not made it into a
> > released Greasemonkey version, for reasons I don't remember
> > (they may not have been aired publicly yet, for instance).
>
> No particular reason, except that nobody has had time to do it. We
> should not be adding random code (no offense) to GM just because
> a patch exists.

Certainly not; code review and working it in the right way is never optional.

Is there anything one can do as a non-committer to help the cause?
(It's a very high up wishlist item of mine, too, since it would reduce
most of my everyday GM scripting code payloads to fewer lines than the
GM script documentation and header itself.)

--
 / Johan Sundström, http://ecmanaut.blogspot.com/
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[WillBo]

The easiest way to do it without having to worry about any complexity and
such would be to just read the library file and insert it into the injected
script right before injection. The only "problem" with this is that the
code would have multiple instances, which is only a "memory space thing".
They would all be sandboxed, so it wouldn't matter that there were multiple
instances of the library code. (not optimal, but it would be easy, and
would work)

Alternately, I was thinking on a way to treat each library as an object,
with
each function a method of that object, with Gm being the (initial) link
between
the scripts and calls to the object methods, since there would be a
timing issue
of when the library was injected itself and available for calls.

So, initially, when you make a call to the object method, you might actually
get "caught" by Gm, which would finish the call once the library code
(object)
was available, and would then modify the calling script's object value
so that
calls to the methods would be direct after that. This may or may not work or
be viable or optimal, I was just mulling it over.

I think with anything other than my first idea, you have to be careful
because
of security concerns and the like, which makes implementation more complex.

Unless someone has already worked out something better that I don't
remember.

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Stefano POGLIANI]

Bill

    how would you do for the first suggestion? Is this "pasting" the
external script code inside the current GM script or is there a coolest
way to do?

I just tried the GM_xmlhttpRequest approach from Philip but in my case
it did not work as the "eval" seems not to return.

Thanks
Best regards
/Stefano

Bill Donnelly wrote:

> The easiest way to do it without having to worry about any complexity and
> such would be to just read the library file and insert it into the
> injected
> script right before injection. The only "problem" with this is that the
> code would have multiple instances, which is only a "memory space thing".
> They would all be sandboxed, so it wouldn't matter that there were
> multiple
> instances of the library code. (not optimal, but it would be easy, and
> would work)
>
> Alternately, I was thinking on a way to treat each library as an
> object, with
> each function a method of that object, with Gm being the (initial)
> link between
> the scripts and calls to the object methods, since there would be a
> timing issue
> of when the library was injected itself and available for calls.
>
> So, initially, when you make a call to the object method, you might
> actually
> get "caught" by Gm, which would finish the call once the library code
> (object)
> was available, and would then modify the calling script's object value
> so that
> calls to the methods would be direct after that. This may or may not
> work or
> be viable or optimal, I was just mulling it over.
>
> I think with anything other than my first idea, you have to be careful
> because
> of security concerns and the like, which makes implementation more
> complex.
>
> Unless someone has already worked out something better that I don't
> remember.
>
> _______________________________________________
> Greasemonkey mailing list
> Greasemonkey@...
> http://mozdev.org/mailman/listinfo/greasemonkey

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[WillBo]

What I'm talking about has to be done down inside Gm and be a supported
feature.
There is no easy or good way to do it, imo, otherwise.


Stefano POGLIANI wrote:
 > Bill
 >
 >    how would you do for the first suggestion? Is this "pasting" the
 > external script code inside the current GM script or is there a coolest
 > way to do?
 >
 > I just tried the GM_xmlhttpRequest approach from Philip but in my case
 > it did not work as the "eval" seems not to return.
 >
 > Thanks
 > Best regards
 > /Stefano
 >
 > Bill Donnelly wrote:
 >> The easiest way to do it without having to worry about any
complexity and
 >> such would be to just read the library file and insert it into the
injected
 >> script right before injection. The only "problem" with this is that the
 >> code would have multiple instances, which is only a "memory space
thing".
 >> They would all be sandboxed, so it wouldn't matter that there were
multiple
 >> instances of the library code. (not optimal, but it would be easy,
and would work)

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[jamescowan]

Is it possible to use venkman to debug a gm user.js script?

James

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Dave Land]

On Aug 11, 2006, at 3:30 AM, James Cowan wrote:

>
> Is it possible to use Venkman to debug a gm user.js script?

Did you ever get an answer to this question?

Dave
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[esquifit-2]

GM scripts are executed in greasemonkey.js, more exactly, in the
evalInSandbox method, which in turns calls the native function
Components.utils.evalInSandbox [1]:

(greasemonkey.js, 223) Components.utils.evalInSandbox(code, sandbox);

Being written in C++, this cannot be debugged with Venkman.

Maybe there is way to slightly modify greasemonkey.js in order to
substitute the native evalInSandbox with something else that can be
debugged?

Incidentally, while searching for information regarding this
possibility, I came across Mozilla Security Advisory 2006-31 [2].  I'm
more than a bit surprised that this, being critical for GM, hasn't
been mentioned in this list till now. This should be discussed in
another thread, though.

[1] http://developer.mozilla.org/en/docs/Components.utils.evalInSandbox
[2] http://www.mozilla.org/security/announce/2006/mfsa2006-31.html

2006/8/21, Dave Land < land@...>:

> On Aug 11, 2006, at 3:30 AM, James Cowan wrote:
>
> >
> > Is it possible to use Venkman to debug a gm  user.js script?
>
> Did you ever get an answer to this question?
>
> Dave
> _______________________________________________
> Greasemonkey mailing list
> Greasemonkey@...
> http://mozdev.org/mailman/listinfo/greasemonkey
>

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Anthony Lieuallen]

On 8/21/2006 3:27 PM, esquifit wrote:
> Incidentally, while searching for information regarding this
> possibility, I came across Mozilla Security Advisory 2006-31 [2].  I'm
> more than a bit surprised that this, being critical for GM, hasn't
> been mentioned in this list till now. This should be discussed in
> another thread, though.
>
> [2] http://www.mozilla.org/security/announce/2006/mfsa2006-31.html

Here's my take.

The linked page specifically mentions that "... a malicious userscript
could gain enough privilege to install malware, but even when
Greasemonkey is working as designed a malicious userscript can make life
miserable. Only install userscripts from sources you can trust."

The purpose of evalInSandbox, for GreaseMonkey, is to separate the user
script from the content page.  If a mozilla bug makes *that* not happen,
then it is a big problem.  The threat of a user script acting malicious
is smaller, if only because the number of user scripts you run is surely
smaller than the number of web sites you visit.

It was also fixed 2 patch levels (minor revisions?) ago =)
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[esquifit-2]

2006/8/21, esquifit <esquifit@...>:
> Incidentally, while searching for information regarding this
> possibility, I came across Mozilla Security Advisory 2006-31 [2].  I'm
> more than a bit surprised that this, being critical for GM, hasn't
> been mentioned in this list till now.

Oh, I see: "Fixed in: Firefox 1.5.0.4".  Sorry. Forget it.
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Jeremy Dunck]

On 8/21/06, esquifit <esquifit@...> wrote:
> 2006/8/21, esquifit <esquifit@...>:
> > Incidentally, while searching for information regarding this
> > possibility, I came across Mozilla Security Advisory 2006-31 [2].  I'm
> > more than a bit surprised that this, being critical for GM, hasn't
> > been mentioned in this list till now.
>
> Oh, I see: "Fixed in: Firefox 1.5.0.4".  Sorry. Forget it.

Well, I noticed it, but only when looking at the release notes for
1.5.0.4.  It's not a moot point-- you have to wonder if there are
other ways of escaping the sandbox-- but it certainly isn't easily
done.
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[jamescowan]

Hi

I am puzzled by javascript scope within greasemonkey/firefox extensions.

In normal web app, I can open a window (wdw = window.open(...) and then
within the parent call the child using wdw.someFunction() or within the
child execute a function in the parent using window.opener.someFunction.

Within a greasemonkey script or firefox extension, I want to open a window
and then make calls from parent and child to functions within the extension.
The standard web technique does not work - I guess because the functions are
not in the scope of the window but of the browser.

Is there any way around this?

James

_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey

[Anthony Lieuallen]

On 10/19/2006 9:07 AM, James Cowan wrote:
> I am puzzled by javascript scope within greasemonkey...
> Is there any way around this?

(Not intentionally rude but...) Learn how greasemonkey works.  A good
place to start:
http://www.oreillynet.com/pub/a/network/2005/11/01/avoid-common-greasemonkey-pitfalls.html
_______________________________________________
Greasemonkey mailing list
Greasemonkey@...
http://mozdev.org/mailman/listinfo/greasemonkey