MapTool Dev Build (testing, version 0.95) - Jan. 28, 2012

[Lee]

@wolph did you mean that it should be like it is in 0.95 where users can toggle it by selecting either day or night? Or do you mean that more menu items should be put in?

When I get the time tonight or some time tomorrow, I'll figure in a nice suggestion from jfj into the handling for day, where a user can clear more parts of the path by setting a cell as a waypoint.

[jfrazierjr]

edit2: for testing purposes I've added a forest. This is a terrible map to walk around regardless of bugs but its good for testing.

Ok.. I am going to go ahead and say it: This is not the use case VBL was made for. 50 trees and and huge vision range(1000) and vision type of circle is going to bring things to a crawl. I can almost guarantee that if you went back to say b75 or so(ie, prior to my changes to add individual FOW) and rebuilt this map(can import it since features in the map did not exist back then!) you would experience the same slowness. One(of many) of the main issues here is the large number of line segments created by all of the squares that have to be merged and add to this the circular vision shape which creates a TON of curves that are a PITA for the java code to compute.

IMHO, VBL was really made for and works extremely well for the use case of rooms and hallways. I will say I have built an extremely large "dungeon complex" when I was working on the code changes for Individual FOW and really can't say that I ever had any type of major issues with speed even moving multiple tokens with individual FOW all around that "dungeon" (and even some outside). Of course, I was using square vision since I know without a doubt that circle vision, while it might be far more realistic, is a huge performance hog(well.. just more far more so that square of the same size.) There is a reason why we say don't use the circle VBL tool...

While it might be fun to block vision for individual trees, that just won't work with the current code approach(and that can't change in 1.3). The right approach will require the MT dev team to stop using the oracle geometry implementation and build our own(probably using some combination of the two in some places.)

In the meantime, the only possible change I can think of that might bring some increased performance in 1.3 would be as noted earlier and that would be to only perform the expose at specific points(such as token "end" point or "end" point + way points) vs each and every cell the token was dragged through. This will cause some loss of functionality, BUT that functionality can be retained 100% by user action(ie, setting of waypoints).

[wolph42]

@lee: I meant an extra toggle, somewhere in that menu
edit: ok thats also not really helping, this hopefully is:
Map menu, between "Fog-of-war" and "Vision"
"Soft Fog-of-war ctr+shift+w"
and/or
"Clear FoW at Waypoints only" (although *that* should probably be a server setting)

@jfj: I'll respond later (currently mobile)

edit:
ok behind a keyboard.

@jfj
First off, your implementation of individual FoW is awesome and lets MT really distinct itself from anything else out there. So both my compliments and thanks for that.
Next: yes the lag was already there before you implemented iFoW, which does *not* make it any less of an issue.
For the rest:
- the vbl used are crosses (not sure that was clear).
- in my example I had iFoW turned off
- nobody ever told me that square vision has performance increase of circular vision (im gonna test that one out)
- vision distance of 1000 or 20 makes (nearly) no difference on that map
- yes the forest *is* overkill and I only put it here for testing purposes, so you can clearly distinct any performance improvements.
That said:
- the forest *shouldn't* be overkill, its not *that* big nor complex and lee proved that with shutting down the soft fow there is 0 lag.
- also its not only this forest, I've had other maps quite large (150x100 cells) with square rooms only, and there the lag also kicked in after some time or even immediately if I had 4 clients attached (usually I have only 1 as I use beamer laptop config). And that *is* an issue.

So given that and the fact that MT functionality *can* remain exactly the same as under b87 but with one extra menu option that drastically improves performance, when needed, at the cost of soft fow functionality only, I sign up for it.

edit3:
tried the square vision and immediately encountered something that maybe is due to my lack of understanding, so if anyone can explain:
If I use square vision: Square Vision: square distance=32,5
and set the map vision distance to 10 then the token can see 32,5 distance. In other words, the map vision distance is ignored when I set the distance. Is that correct? I would expect the map vision distance to be leading at all time, but of course I could be wrong.

[Sol Invictus]

- also its not only this forest, I've had other maps quite large (150x100 cells) with square rooms only, and there the lag also kicked in after some time or even immediately if I had 4 clients attached

yup, here too. Even on maps not that big and with very limited VBL. 4 clients are minumum in the games I run, and that used to cause horrible lags with b87. Haven't tried the dev build in such an environment yet.

[RG_kortor]

Did anyone get it to run under Linux?

i realize this thread may have become archived, but yes i have run maptools in linux.
i have not yet got to work under openjre6 or 7, but if you install the windows oracle JRE SE binary via wine, and move the maptools folder into the .wine/drive_c area, you can open the .jar via Java Binary SE.

so far it has worked flawlessly for me, granted i am still running 1.3b63

[wolph42]

I think he meant the current dev version. Cn you try that?

[jfrazierjr]

Did anyone get it to run under Linux?

i realize this thread may have become archived, but yes i have run maptools in linux.
i have not yet got to work under openjre6 or 7, but if you install the windows oracle JRE SE binary via wine, and move the maptools folder into the .wine/drive_c area, you can open the .jar via Java Binary SE.

Thanks for the input, but unless you have some philosophical issue, why not just install the native oracle client on linux instead of dealing with the wine? Granted, you have to actually go looking for binary install files, but IMHO, that's better than the additional overhead emulation deals, especially since there is an alternative.

[Lee]

I just wanted to provide an update as I've been made to start early at work. I haven't been able to start putting in jfj's suggestion but I'll do so once I can find some free time.

[RPTroll]

I think he meant the current dev version. Cn you try that?

Yes. I downloaded and was unable to get it to start. Granted, I didn't try very hard.

[Lee]

Edited (had xdg-utils installed all along and it didn't work -> needs tweaking)
@RPTroll Since there's bound to be a demand for an open Java console window (i.e. via java), I made a change in the Start-MapTool launcher between version 0.9 to 0.95 to route the java call through the OS shell. I found that using xdg doesn't get the desired results, I might just write a script and call it from the launcher. Let me work on that when I get the time.

If the java console isn't important to you and you still have version 0.9, you can use the 0.9 launcher as it makes a direct Java call at the loss of the console window. I can confirm, however, that 0.95's maptool-dev.jar will launch when run from the terminal, at least from what I've tested in Ubuntu. 0.9 launched as expected using the launcher, FWIW.

I would have done so had I had more time, but moving forward, I'll put a logical branch in to go through the OS shell when the user selects the "Console" checkbox and make a direct Java call when it isn't (I had finished compiling and uploading 0.95 when I realized I forgot to put in, sorry).

[RG_kortor]

Did anyone get it to run under Linux?

i realize this thread may have become archived, but yes i have run maptools in linux.
i have not yet got to work under openjre6 or 7, but if you install the windows oracle JRE SE binary via wine, and move the maptools folder into the .wine/drive_c area, you can open the .jar via Java Binary SE.

Thanks for the input, but unless you have some philosophical issue, why not just install the native oracle client on linux instead of dealing with the wine? Granted, you have to actually go looking for binary install files, but IMHO, that's better than the additional overhead emulation deals, especially since there is an alternative.

the overhead for this application isn't enough to really cause me any issues. and it isnt due to any real philisophical holds. I am trying really hard to convert to linux from windows (just getting tired of microsoft's view of computing). finding the binaries wasn't difficult, and with google search, i had at least half a dozen howto hits. but they were evidently incomplete, or the process just didn't work for me. so the wine work around was quick, clean and convenient.
the emulation "lag" running maptools as a virtual tabletop for a PnP DnD isn't nearly as noticeable as it is on a FPS MMO, so this solution is suiting me just fine although it is not what everyone needs.

[jfrazierjr]

the overhead for this application isn't enough to really cause me any issues. and it isnt due to any real philisophical holds. I am trying really hard to convert to linux from windows (just getting tired of microsoft's view of computing). finding the binaries wasn't difficult, and with google search, i had at least half a dozen howto hits. but they were evidently incomplete, or the process just didn't work for me. so the wine work around was quick, clean and convenient.
the emulation "lag" running maptools as a virtual tabletop for a PnP DnD isn't nearly as noticeable as it is on a FPS MMO, so this solution is suiting me just fine although it is not what everyone needs.

That's cool. while in a general sense it's probably a tiny amount of lag, if you really push the boundaries of MT's graphic capabilities, it might be more noticeable(complex vision blocking, lighting, etc). Granted, even on a good fast computer, throwing to much at MT can cause lag, so I would personally just prefer to keep that footprint as small as possible, but to each his own....

[Sol Invictus]

Hi again! Today's gaming session we played online with the dev build. I got a player of mine to agree to test in detail with his Linux system, I will report on that when we actually did that in detail.

Everything went fine today and as expected on my end, and for most players, but not for all, unfortunately. One of my players ran this macro:

macro

Code: Select all

Fire Burst<br>[h:numOfAttacks=number][count(numOfAttacks,"<br>"), CODE:{Attack [r:roll.count+1]: [t:1d20 +11] vs. Reflex}]<br>Damage: [t:3d6+8]<br>

What we all got as chat output was weird:

chat output

CharacterBG3_02:
Fire Sea Travel
Attack 1: 21 vs. Reflex
Attack 2: 15 vs. Reflex
Attack 3: 23 vs. Reflex
Damage:Attack 1: 21 vs. Reflex
Attack 2: 15 vs. Reflex
Attack 3: 23 vs. Reflex
Damage:vs. Reflex
Attack 2: 15 vs. Reflex
Attack 3: 23 vs. Reflex
Damage:Attack 2: 15 vs. Reflex
Attack 3: 23 vs. Reflex
Damage:vs. Reflex
Attack 3: 23 vs. Reflex
Damage:Attack 3: 23 vs. Reflex
Damage:vs. Reflex
Damage:Damage: 15

The output should look like this:

output as supposed to be

CharacterBG3_02:
Fire Sea Travel
Attack 1: 30 vs. Reflex
Attack 2: 30 vs. Reflex
Attack 3: 16 vs. Reflex
Damage: 15

And it does look like that, i.e. correct, when the macro is run from another machine. The output also is correct when the macro is run on his machine in b87 and b88. We checked and Java 6 or 7 do not show any differences.
This is the debug info for his machine (win 7, 64bit):

Spoiler

==== MapTool Information ====
MapTool Version: DEVELOPMENT
MapTool Home...: C:\Users\Derp\.maptool
Max mem avail..: 989M
Max mem used...: 62M

==== Java Information ====
Java Vendor.: Oracle Corporation
Java Home...: C:\Program Files (x86)\Java\jre7
Java Version: 1.7.0_11
............: Java(TM) SE Runtime Environment (build 1.7.0_11-b21)
............: Java HotSpot(TM) Client VM (build 23.6-b04, mixed mode, sharing)


==== OS Information ====
OS Name........: Windows 7
OS Version.....: 6.1
OS Architecture: x86
Processor......: Intel64 Family 6 Model 30 Stepping 5, GenuineIntel
PATH...........: null
Number of Procs: 8

==== User Information ====
User Name: Derp
User Home: C:\Users\Derp
User Dir.: C:\Users\Derp\Downloads\D&D\Mein Char Ver 4.0\maptoolsneu\MT BETA 0_9_5

==== Network Interfaces ====
Display Name..: Software Loopback Interface 1
Interface Name: lo
Address...: 127.0.0.1
Address...: 0:0:0:0:0:0:0:1

Display Name..: Intel(R) 82577LM Gigabit Network Connection
Interface Name: eth3
Address...: 192.168.1.34
Address...: fe80:0:0:0:4135:9dd3:5847:5ca5%10

Display Name..: Teredo Tunneling Pseudo-Interface
Interface Name: net5
Address...: 2001:0:5ef5:79fb12bdfedd
Address...: fe80:0:0:012bdfedd%14

Host Address...: 192.168.1.34
Default Gateway: 192.168.1.1

==== Locale Information ====
Country.: Deutschland
Language: Deutsch
Locale..: Deutsch (Deutschland)
Variant.:

==== Display Information ====
Number of Displays: 1
Display 1: 1920x1080(32)

==== Internet Gateway Devices ====
Device Name.: ADBB router
Model Name..: ADBB router
Manufacturer: ADBB
Model Number: 1
Model Desc..: ADBB router
Firmware....: EPICENTRO UPnP/1.0 MiniUPnPd/1.6
External IP.: 92.107.232.147

The creator of the application that created this macro pinned the problem down to <br> tags (he's registered here as chillhelm I believe). <br> tags in macros that follow dice rolls will cause the following text to be repeated.
That's his example:

Code: Select all

blabla 
<br> roll [1d8] <br>
this text will be repeated when run by on this particular computer in the dev build
[1d6]
<br>

edit: Translated the chat output to make it more readable.

[Natha]

Did yo try with <br/> instead of <br> ?

[aliasmask]

Does that user have "Use tooltips for inline rolls" checked in preferences? I would also recommend using [r,count... for loop.