[Plugin] VisMap - Layer/Scene Visibility Map

takesh h

Hi Martin,
A tour de force.
I tried it on OSX/SU7 and it worked beautifully. Didn't work on SU6 though (got a blank window).
The problem I saw was that it doesn't stay in front if you go to another program and come back to SU.
It just hides itself in the back of SU, for SU window is usually bigger than vizmap window.
Other than that, it worked flawlessly as if it's designed for OSX!

EDIT: I even included a couple of layers/scenes with Japanese names and it seems to have no problem displaying them.
On OSX/SU7, vismap window appears to be blank right after it's opened, resizing it or clicking on the window doesn't make the situation any better. I just had to go to another program once and come back to see it is actually working properly.
I thought it was just taking a long time to display what it should be displaying, but it doesn't seem to be the case.

MALAISE

Hi Martin

Very useful overview to decide which layers have to be displayed on which scenes .
I just got a strange message several times on one model ( I may upload it if needed)
Choosing OUi or NON is without effect.The vizmap leaves blank.
I hope it can help .

MALAISE

some translations : "Attendu" means waited
" Voulez-vous continuer à executer les scripts sur cette page : Continue to run the scripts of this page ?
" Une erreur est survenue sur le script de cette page " an error occured when running the script ...

Edson

i tried it on OSX/SU 7 but thw vismap window appears blank no matter what i do. i opened several files but the problem remained.

markozeta

Naming your scenes BM6.2 or MS4.8 (Which is the official name) ... will kill the program.

Sighs

Also another request for those with complex models is the ability to use freeze panes on the scene list and layer list.

Thanks for the work though, your script will help us put everything into layout correctly.

MartinRinehart

@takesh h said:

I tried it on OSX/SU7 and it worked beautifully. ... it worked flawlessly as if it's designed for OSX!

@edson said:

i tried it on OSX/SU 7 but thw vismap window appears blank no matter what i do. i opened several files but the problem remained.

Help!

Edson, I'm adding some bullet-proofing that will, in some cases, tell you where it's failing. This may be constructive.

@markozeta said:

Naming your scenes BM6.2 or MS4.8 (Which is the official name) ... will kill the program.

Ahhhh! The sort of problem that might be solvable. Thanks.

@markozeta said:

Also another request for those with complex models is the ability to use freeze panes on the scene list and layer list.

Freeze panes? What is a freeze pane?

Martin

MartinRinehart

@malaise said:

I just got a strange message several times on one model ( I may upload it if needed) ... The vizmap leaves blank.

Yes, the model (ideally the subset that causes the problem) would be key to tracking this down.

MartinRinehart

@takesh h said:

vismap window appears to be blank right after it's opened, resizing it or clicking on the window doesn't make the situation any better

Working on it. Be patient. The U.S. Open is in week two, which really cuts into my free time.

Very, very odd: your screenshot shows buttons on top, icon and title on bottom, other things reversed. As a guess your appendChild() JavaScript method is appending to the front of the list while my PC appendChild() appends to the end of the list. Hmmmm.

Edson

please let us know when you post a modified version.

markozeta

@martinrinehart said:

@markozeta said:

Also another request for those with complex models is the ability to use freeze panes on the scene list and layer list.

Freeze panes? What is a freeze pane?

Martin

Since your a Linux man, I'll quote you the OpenOffice version instead of the excel version:

http://inpics.net/tutorials/calc2/vis5.html

It would help a ton, but most code that you could drop into your javascript routine would try to freeze the top and side, when you want the two sides to be frozen.

MartinRinehart

@edson said:

please let us know when you post a modified version.

There is a new version now fixing a Mac issue. The strange "backwards UI" thing should now be gone.

Does this have any effect on Edson's "no UI" issue? I doubt it, but you never know.

Edson

@martinrinehart said:

Does this have any effect on Edson's "no UI" issue? I doubt it, but you never know.

i am afraid there has been no change: still no UI.

MartinRinehart

@markozeta said:

Naming your scenes BM6.2 or MS4.8 (Which is the official name) ... will kill the program.

I changed "ground_floor" and "second_floor" to "ground.floor" and "second.floor" without problem. Can you post a file? Thx.

markozeta

Martin,

The issue seems to be size. I could give you the model, but it's too large for these forums to post! However, as I was trying to reduce the size of the model in order to post it, when it got under 3.4 MB or so, the script worked!

The bug seems to be not in getting data, but in sending it back to the model. When I do that, I get a prompt asking me if I want to continue running scripts.

This prompt states "Error at Line 78, Char 4" of the script, and the error given was "The data area passed to a system call is too small."

Could this be an overflow error of some type?

Edit: I have 50 layers and 43 scenes, if that helps. That's a total of 2,150 check points. When deleting to fit to size, I also "purged unused", and then it worked.

Mark

MartinRinehart

@markozeta said:

Martin,

"Error at Line 78, Char 4" of the script, and the error given was "The data area passed to a system call is too small."

Could this be an overflow error of some type?

Edit: I have 50 layers and 43 scenes, if that helps. That's a total of 2,150 check points.

Thanks for your help!

Line 78, char 4 is the middle of a variable name. MSIE's error reporting is very near to totally worthless. It has to be SketchUp passing data to a system call, so this is a SketchUp bug. (Remember that numbers like 50, 43 and 2150 are very small numbers - though the VisMap you get probably looks huge.)

Today I am going to try to get good with one or the other of Ruby's alternative exception handlers and see if I can get enough information to pass back to the JavaScript so you get an error message that might be helpful.

MartinRinehart

@markozeta said:

[Freeze panes] would help a ton, but most code that you could drop into your javascript routine would try to freeze the top and side, when you want the two sides to be frozen.

OK. I'm up to speed on freeze panes. What exactly is your design?

MartinRinehart

I just ULd new Ruby and JavaScript.

The Ruby now traps errors in the bit of the code where I think they might happen, passes the error message on to the JavaScript. In turn, the JS uses the error message to create an error report where you were expecting the VisMap UI.

If my guess is right about the source of the trouble, there should be no more blank UIs.

Edson

martin,
there has been some improvement: the UI is no longer blank, as you can see, but it does not get any data from the model. i have 24 scenes and about 40 layers in this model. i am on mac (10.5.8/SU 7)

MartinRinehart

@edson said:

there has been some improvement: the UI is no longer blank

Beautiful!

First, I see that your scene list is on the wrong side. I can fix that. (Edit: done.) Second, it appears that there was an issue with your scene list, as it appears to have broken before/during the scene name collection process. I'll cast a wider net for breaks and see if I can get you a genuine error message. (Edit 2: also done.)

(It's rained all day. The U.S. Open is closed.)

Edson

martin,

i downloaded the plugin again but not only there was no improvement as we are back to the blank UI. keep trying.

markozeta

Aha! It is a bug with overflow.

http://www.boutell.com/newfaq/misc/urllength.html states that IE can only have a URL of length 2,083 chars.

Line 78 to 82 of vismap.js states...


function rubyCalled( callback_name, message ) {
	if ( (typeof message) == 'undefined' ) message = '';
	url = 'skp;' + callback_name + '@' + message;
  	window.location.href = url;
}

With my example, this would request a url from internet explorer that would contain a 2,150 character string attached to an 11 character header to form a 2,161 character url - which exceeds IE's ( ) limits.

I guess the only way I could suggest fixing this is with a conversion of binary (1 for V, 0 for I), converted to ASCII, and then back on the ruby side. Ruby will be able to unpack it though, using str.unpack in ruby, but still, this will be difficult to overcome.

.>

Not much help, sorry.