WebDialogs - The Lost Manual — R1 09 November 2009

chrisglasier

Finally, there is a difference in the way that the Mac boots up SketchUp that you should be cautious about: there is no Sketchup.active_model when the Ruby scripts are first loaded. So if your script is making changes to the active_model at load time, it will not work on the Mac. The answer? Ensure code that references the active model is part of a UI event handler, responding to the user selecting a tool or a menu item. You can also use an AppObserver to get a callback whenever a new model is opened, at which point it's safe to talk to the active_model.

I didn't see this anywhere. I don't remember where it comes from.

How about marking topics and posts with [WebDialog]?

Good work - most generous with your time.

Chris

thomthom

I hope that people will point out things that are not clear. Writing has never been my strongest skill, actually - its one of my poor, so please point out the weaknesses.

tomasz

Thank you!

Jernej Vidmar

Hi Thom,

thanks for sharing this really usefull information!

Just as a side note:
We have found a bug, which cuts decimal value sent from SketchUp to WebDialog (only in SU6 on Mac). If we i.e. want to update the value of some input box like this:
web_dialog.execute_script("updateInputBox('varName', '12,345')")
only 12 is shown in the varnName inputbox, everything behind (and including) comma is being cut off. But it works in SU7 on Mac and on both, 6 and 7 versions of SU on Windows.

thomthom

Ah, is that whatæs being mentioned in the release notes: http://code.google.com/apis/sketchup/docs/releases.html

@unknownuser said:

Fixed Mac support for WebDialogs execute_script

WebDialog.execute_script('alert("Bug is Fixed!")');

Does it work if you escape the comma? If not, replace it with another character?

Jernej Vidmar

No, escaping doesn't seem to help. Only way to pass the float value is to write it using dot (1.12) not comma (1,12)...
For now we have not found a workaround, but we intend to make some more tests to see if the problem can be bypassed.

thomthom

I wonder, if you make a special receiving javascript function, and send all command base64 encoded to that function and have it decode and eval it..?


# made up methods - haven't checked the real methods
jscall = base64encode("updateInputBox('varName', '12,345')") 
WebDialog.execute_script("decode(#{jscall})");


function decode(base64str)
  eval(decode64(base64str));
end

DIEGO-RODRIGUEZ

interesting applications for html

html 5
http://html5demos.com/

and

http://www.dhtmlgoodies.com/index.html?page=dhtml-suite

thomthom

You won't be able to use HTML5 in webdialogs - not until IE adds support for it. Think I read somewhere that they where adding support for it in IE9. Being able to use CANVAS sure would be nice.

tbleicher

@thomthom said:

You won't be able to use HTML5 in webdialogs - not until IE adds support for it.

I am using excanvas http://code.google.com/p/explorercanvas/ which provides a canvas object in IE.

Thomas

thomthom

@tbleicher said:

@thomthom said:

You won't be able to use HTML5 in webdialogs - not until IE adds support for it.

I am using excanvas http://code.google.com/p/explorercanvas/ which provides a canvas object in IE.

Thomas

That's very interesting! I've been wanting for a good solution to dynamically draw graphics in webdialogs. Probably be of interest for Whaat as well... ...thinking Profile Builder...

Jim

And just to have more options, there is also chrome frame.

thomthom

@jim said:

And just to have more options, there is also chrome frame.

Yea, but this requires the user to install a browser plugin.
explorercanvas is a simple JS library which the developer includes in the project without the user ever having to install any browser extension. That's what's appeal to me.

Pout

this seems to be very interesting
Have to dig into this.

chrisglasier

One thing driven discovered was that if for some reason there is whitespace as here -

cmd = "fImportReturn('#{array}');"
@dlg.execute_script (cmd)

instead of

@dlg.execute_script(cmd)

it is OK on PCs but not Macs. This maybe of interest to coders who use this style (Jim).

Jim

Would a kind Mac user (or 2) open the Ruby Console, and give the result from entering:

$VERBOSE

and

$DEBUG

and

$-w

thomthom

@jim said:

Would a kind Mac user (or 2) open the Ruby Console, and give the result from entering:

$VERBOSE

and

$DEBUG

and

$-w


> $VERBOSE
false
> $DEBUG
false
> $-w
false

SU 7.1.6859
OSX 10.4

Dan Rathbun

By the way.. we have no control over C-implemented Ruby objects, that incorrectly call the wrong C 'warn'/'warning' function, ie: don't respond to the setting of $VERBOSE ( called ruby_verbose in C.)

If you find one, it would need to be reported on RubyForge.net

chrisglasier

@mmyoung said:

I was never able to get Matt666's PointTool.rb to work on my Mac. I opened it in a text editor and removed the "=begin" and "=end" lines at the beginning and placed a "#" at the beginning of each comment.

Now it runs on Mac.

This may be relevant here as well.

driven

@chrisglasier said:

This may be relevant here as well.

hi,
some Mac bits...

@Jim I get false, false, false

@Dan #!ruby warn_ovr.rb returns nil

whitespace handling is strange, not consistent from script to script, some work with whitespace , but all work without it...

the =begin; =end seems to be dependent on what editor the file was last saved in....

I never had a problem if I had a look with xCode, but when I first used Smultron the syntax colouring for '=' commenting is off by default, and very oddly appears to effect how SU then sees some of the files saved from this state...

It took a while to figure out, but all affected files worked after changing the syntax colouring option to on and re-saving. Re-saving from xCode also 'fixed' them, which was my only 'clue' of what was happening.

Something that came up proof positive last week (and I'm trying to find a test for), is that not all Macs are handle syntax the same

there was an issue with one of TIG's scripts and it seems that Mac's running Developer Tools have much stricter syntax prerequisite.

I did a mini survey of systems 'failing' and we all had different mixes of OSX, Ruby, Gems, editors but all had DevTools

a Variety of Macs without were fine but those with Xcode were failing to run the script, TIG found a solution for his script, but the consistency of the Pass/Fail leads me to think it may make a good test tool... when I know how...

On another point anyone know of a .json to/from .skp or collada convertor for Mac?

john