MapTool 1.3 Development Build 58

[jay]

@jay: was the initiative panel open?

Yes it was.

[Micco]

Contributed by Mark Karasek
* Updates:
[s: ] - only display to self
[g: ] - only display to GM
[w("name"): ] - only display to specific player
[st: ] - only display tooltip to self
[gt: ] - only display tooltip to GM

These can be combined, so, for example, the output of [s,g,st: ] would appear to yourself and the GM but not other players, and only you would have a tooltip for it.

I haven't used it yet, but I was scrolling through the posts here to say THANK YOU! This is awesome. Like a little early Christmas present

I'll test it on some macros tonight and post if there's anything funny going on.

[plothos]

I was scrolling through the posts here to say THANK YOU! This is awesome. Like a little early Christmas present

Ditto in spades.
My NPC attack macros especially thank you. (They're really very polite little macros.)

[Azhrei]

I got this message in the console, but nothing in the log.txt file:

Code: Select all

Sep 22, 2009 10:13:53 AM com.jidesoft.docking.DefaultDockingManager a
WARNING: Error loading layout - null

Yeah, I don't think we can trap the JIDE errors and redirect them to the log file.

But this error is pretty clear what the problem is: the layout.dat file that saves the layout can't be read back in. I'm not sure how the JIDE library does its I/O. It's definitely a binary format and may just be object serialization. But I can single step through main() and it's the point at which the layout is read in where things go south.

It may be that an exception is being thrown as the layout is being written out to the file and some layer of code catches the exception but doesn't handle it properly.

[Azhrei]

I have modified the ZIP file that Trevor uploaded for b58 so that the execute permission is turned on for the Unix/OSX shell scripts. This should make it much easier for those users -- just double-click the file in the GUI to start it.

[trevor]

I intended a build last night, but sorta fell asleep at my desk while looking into the layout issue.

I updated the docking framework jars to the latest, but that didn't fix it. So I'll need to search through the JIDE support forums.

Looking through the forum posts, this has been happening for a while, but seems to be getting progressively worse (even though I'm not directly touching anything related to it, bizarre).

There will be a built no later than tonight, with the layout fix or not.

[lmarkus001]

I've removed the big nasty warning message from b57. Things have seemed pretty stable.

Thanks for the update, as always it is greatly appreciated and excitedly awaited!

One comment on your note. b57 was so buggy I was unable to use it to run my campaign. It is the first release that was ever too buggy to run. I mention this as b58/9 might uncover a few more hiccups (hopefully not though as the fix list looks pretty good).

[trevor]

b57 was so buggy I was unable to use it to run my campaign.

Hmm, did b58 fix it ?

[lmarkus001]

At first glance it looks all better. I will use b58/9 this Friday in my active campaign to give it a good workout.

[trevor]

Good News !

The fantastic people at JIDE software have really come through.

I posted the issue in their support forums and within a few minutes had a response, and within a couple of interchanges, got a pre release version that prints out the full stack trace of the problem.

So I was able to figure out _exactly_ what was causing the layout issue. So the initpanel issue is fixed, and I'm going to go fix the other panels too.

Definitely a new build tonight

[RPMiller]

Awesome Trevor!

[Azhrei]

So I was able to figure out _exactly_ what was causing the layout issue. So the initpanel issue is fixed, and I'm going to go fix the other panels too.

I'm looking forward to checking out the code to see what the changes are! It sounds like it involves changes to MT as well as an updated library from them...

[trevor]

It sounds like it involves changes to MT as well as an updated library from them...

There were a couple problems:

- JIDE docking was initializing the frames as it laid them out
- This was happening before the panels (init, and tables) were being initialized
- The init and table frames throw an exception if they are laid out before being initialized
- JIDE was swallowing the actual error and printing out the useless WARNING line

So the prerelease version they gave me printed out the full stack, which immediately identified the failure in the panel inits.

The resulting fixes were quite trivial.

[thecyberwolfe]

Proving once again that knowledge is power

[Azhrei]

There were a couple problems:

- JIDE docking was initializing the frames as it laid them out
- This was happening before the panels (init, and tables) were being initialized
- The init and table frames throw an exception if they are laid out before being initialized
- JIDE was swallowing the actual error and printing out the useless WARNING line

Yeah, that last one is a particularly annoying habit of some libraries. I understand why they do it in the general case, but it makes tracking down problems insanely difficult.

Um, that's more than "a couple", btw.

The resulting fixes were quite trivial.

Excellent. Thanks for the info.