[DONE] OD Windows binaries

[Danimal]

Exactly the same happens to me, but i still dont have the fixed version you devs use.

[paul424]

Can we use IRC ?
EDIT ": Updated the FAQ file , give some feedback if you get Camera movemnt now.

[Bertram]

Hi paul424,

I saw the FAQ update. thanks!
I didn't try yet but I sure will.

I just have a few questions, as usual. If you have to use the mouse to rotate and scroll (along with other keys), what are page up, page down, home and end keyboard keys about?
Why would I need to hold an additional key to zoom in/out with the mouse middle button? Isn't it useless or is there another functionality using it?

Best regards,

[Bertram]

Hi again,

EDIT ": Updated the FAQ file , give some feedback if you get Camera movemnt now.

Tested on Win7:
General note: The starting field of view is now 3/4 perspective and not a full top-down one anymore.

The camera panning (left, up, down, right) is still working with mouse and keyboard with latest changes.
I just noticed that sometimes releasing a directional key and pushing another one was making the camera stops. But it's only a minor annoyance.

The camera height, using CTRL + mouse wheel is functioning fine.

Dragging LMB and going to the left or right edge seems to be working basically but it's way too sensitive. Dragging to the right leads almost instantly to a top-down view.
Dragging on the left gives a black screen and the camera seems then blocked or something. Sometimes, it seems the camera went "below" the world.

Again, I'd request to remove the need of CTRL, except if something else is planned for the mouse wheel.
I would also request to remove the LMB dragging method in favour of keys, like page up and page down to tilt the view and use A and E to rotate the camera.
But overall, it's an improvement. Congrats!

On a dev side of things: I've removed the CMakeLists.txt revert from your patch on my repository. Could you give me your cmake and gcc version? I'd like to find a proper fix to the cmake file instead of reverting.

Also, you should beware of trailing spaces and dead code when committing. Except if you did it on purpose? (as for the dead code, of course).

Best regards,

[paul424]

{l Code}: {l Select All Code}

 cmake --version
cmake version 2.8.10.2

 gcc --version
gcc (SUSE Linux) 4.8.1 20130909 [gcc-4_8-branch revision 202388]
Copyright (C) 2013 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.


Read my PM ...

[Bertram]

Hi paul424,

Seems quite up-to-date. Weird. I think I'll come back to this later. Let's move on with more urgent topics for I don't want to break your dev environment (and mine) any further for now.

Read my PM ...

Did it.

I do think you shouldn't worry too much as an average person check his/her mail every few days, or even every weeks only, but if you have got the mail of the people you want to contact, or can PM them, it doesn't hurt, of course.

Also, IMHO, on the artistic side of things, the only major point that needs progress is upgrading the shaders for certain models, right?
--> I guess I'll have a go at opening the source files in blender and see what can be done, but I can't promise anything done on it soon. Anyone welcome on this task.
Also, I do think, an another problem we have is the missing tiles if there are still any? Could you tell me a bit more about the culling manager and why it is disabled?

Now, I won't hide that the main work to be done is on the game playability and code cleanliness. If we can get to the point where we have a clean basic set of working game commands for the next release, and maybe set up a quick HP system, so we can start dropping creatures on enemies, it will be won.
I'll be working on adding an app icon, and get the code warnings to zero, from time to time.

Again, to make things clearer for everyone, we must get a decision about whether my latest commits can be merged, and whether everyone is ok about having the mandatory media files in the same repository, ready to go, for the sake of simplicity. (The media sources can remain in svn, of course.)
For now, my repository is looking like some kind of integration branch, something that belong more to the official git master (or development) branch.

Best regards,

[paul424]

Look at the dictionary.txt there are all commands listed , some of them could hang the game though ...
For what I speak , please try startcreatureculling starttileculling also try to change camera view angle before or you would get crash.
( naah sorry ) ;(

[Bertram]

Ok, I'll have a look. (But likely not before the end of the week.)

[Bertram]

( naah sorry ) ;(

Err, for what?

[Bertram]

Hi there,

I started removing warnings. could you try this patch and tell me whether I can integrate it?
https://github.com/Bertram25/OpenDungeo ... 9a08a695bc

Best regards,

[paul424]

Looks like having two servers is too much , why you don't create a separate branch at our remotes/origin like Bertram_fork or something , than I would just rebase or merge into development b-ch ....

[Bertram]

Sure Will do it asap.

[paul424]

Still no new branch .

[Bertram]

Be patient, everything happens eventually

https://sourceforge.net/p/opendungeons/ ... tion/tree/

[Bertram]

Hi again,

I've added another commit to the sf branch putting the warnings to zero when using gcc. Please try it also and merge it if it's running ok.

And more importantly, please, PLEASE, stop committing dead code, code that triggers warnings, and messed-up indentations, of arbitrary placed spaces, as I could see all over the current code base.
I'll be honest, the code must be a nightmare to analyse when working on it, since it's a real mess. Let's try to work as a team from now on, and stop adding more to the problem...

Thanks.

[paul424]

Hi, What is a " dead code " for you ?

This is a huge problem ! That there is no intendation standard , no indent scripts , oln claims that he lost the scritpt auto-fromating to standard there is ....
If you are good with those balls you could propose some formatting standard and reformat the source to it.

[Bertram]

Hi paul424,

Thanks for answering my call.

Dead code for me is either code that doesn't get compiled/is commented while having no obvious reason to be there or code that is compiled but never run, again with no actual reasons to stay.
What I'm begging at, is not to commit commented code most importantly. Or at least without any comment on about why it is commented but still here.

I'm also begging at not to commit code that triggers warnings or doesn't compile/run on your system.

Btw, there is a convention, described here:
http://opendungeons.sourceforge.net/w/i ... Guidelines
Mainly, I would ask to go and stick with 4 spaces as an indentation unit but there is also the fact that in many places of the code, spaces for comma separated list are used in a random way. e.g.:

{l Code}: {l Select All Code}

class ::       function_name    (     int x ,      int y     )  ;

where I'd simply request this:

{l Code}: {l Select All Code}

class::function_name(int x, int y);

If indentation and such things could be kept in mind while coding, we're off of 90% of the problem. For the rest, there is AStyle.

Last point, I'd ask not to use AStyle too early as I would like to fix files per files, while looking at improving the code style and/or make sure all class members are initialized properly.
This will help me get into the project, in fact.

But later on, we can set up a Astyle script without any problems.

Best regards,

[Bertram]

Btw, I've found that the mouse wheel is used to change the dragged creature order when not holding CTRL.
It would be me, I'd change this to change the held creatures order when dragging CTRL, and move up/down when not,
as moving is a more basic movement than changing creatures order in hand.

Would you be ok with it?

[paul424]

Hang on #opendungeons sometimes , please ...
I don't stick to details so much , what I want is to complete the XML parser , but sitll have some problems with SAX, I could push of what I have if you feel strong in checking out new code with libxml++ ...

[Bertram]

Hi paul,

Hang on #opendungeons sometimes , please ...

Sure. We can set up a meeting. Just tell me when it could fit for you, and we'll have one of those happy dev times.

I don't stick to details so much

I'll take this for a yes about the mouse wheel change, and commit that along with the rest in my branch.

what I want is to complete the XML parser , but sitll have some problems with SAX, I could push of what I have if you feel strong in checking out new code with libxml++ ...

I should be ok with libxml++. Yet, I will have to know what you're trying to do, to get the idea of what you're coding. For a starter, what are you using the xml parser for and what is currently lacking?
EDIT: Could you push all that in your own branch, so that we can do each other review before pushing in the official one?
This way, we could both avoid pushing broken stuff without knowing.

Now, my turn to present my latest commits, all available in my integration branch

Started removing compilation warnings.
Put the gcc compilation warnings to zero.
--> Fixed warnings on gcc, all except one in angelscript files, which a bit trickier. Once those two will be accepted, we should not commit things that triggers new ones.

MEDIA: Fixed the logo and button behaviour for every screen resolutions.
--> The new GUI was mishandled when not in resolution of 800x600. Now, the buttons stay on center and the logo doesn't stretch anymore.

Put saner height bounds for the camera view.
--> I put a upper bound when changing the view height (z value) and tweaked the min one while on it to fit game better.

Tweaked a bit how debug info is displayed.
--> The green debug text seen when playing is now shown/toggled only when pushing F11. I also made it display loading... until the world and the creatures are displayed, so that the user knows the game is still working.

Best regards,

[paul424]

Bertram !
Now git pull takes forever , I get only 100KB/s from my ISP.
Maybe in France it is diffrent where high speed internet is common .
Imagine new people pissed off who just want to see the code ...
For creating a separete branch for my >> works << there is no problem ... When the current version is very polished keep in mind to refresh the "master" branch as well .

This comes from the Cmake-gui CMakeList.txt :

{l Code}: {l Select All Code}

-std=c++11 -Wall -Wunused -Wno-deprecated -fno-strict-aliasing -std=c++11 -Wall -Wunused -Wno-deprecated -fno-strict-aliasing-std=c++11 -Wall -Wunused -Wno-deprecated -fno-strict-aliasing -std=c++11 -Wall -Wunused -Wno-deprecated -fno-strict-aliasing

Why it is so many times repeated ?

{l Code}: {l Select All Code}

[  0%] [  1%] [  2%] Building CXX object CMakeFiles/OpenDungeons.bin.dir/dependencies/angelscript/angelscript/source/as_atomic.cpp.o
Building CXX object CMakeFiles/OpenDungeons.bin.dir/dependencies/angelscript/angelscript/source/as_builder.cpp.o
Building CXX object CMakeFiles/OpenDungeons.bin.dir/dependencies/angelscript/angelscript/source/as_bytecode.cpp.o
g++: fatal error: no input files
compilation terminated.
g++: fatal error: no input files
compilation terminated.
/bin/sh: -std=c++11: command not found
make[2]: *** [CMakeFiles/OpenDungeons.bin.dir/dependencies/angelscript/angelscript/source/as_atomic.cpp.o] Error 127
make[2]: *** Waiting for unfinished jobs....
/bin/sh: -std=c++11: command not found
make[2]: *** [CMakeFiles/OpenDungeons.bin.dir/dependencies/angelscript/angelscript/source/as_builder.cpp.o] Error 127
g++: fatal error: no input files
compilation terminated.
/bin/sh: -std=c++11: command not found
make[2]: *** [CMakeFiles/OpenDungeons.bin.dir/dependencies/angelscript/angelscript/source/as_bytecode.cpp.o] Error 127
make[1]: *** [CMakeFiles/OpenDungeons.bin.dir/all] Error 2
make: *** [all] Error 2


Your CMakeList.txt seems to be utterly broken , and gernerating invalid Makefiles as one can see above ...

IRC <silly smille >

[Bertram]

Hey,

Bertram !
Now git pull takes forever , I get only 100KB/s from my ISP.
Maybe in France it is diffrent where high speed internet is common .
Imagine new people pissed off who just want to see the code ...

Well, in France, higher internet speed is indeed more common but not at my home. I'm just used to do something else while pulling/pushing.

From personal experience, people are more pissed off when the code is ugly and can't run because of the lack of game data,
rather than waiting once for the initial cloning time.
So I'm definitely sure to be on the good trend.

For creating a separete branch for my >> works << there is no problem ...

Great,

When the current version is very polished keep in mind to refresh the "master" branch as well .

Right. If you're ok, once the code will be reindented and the code style a bit reviewed in the development branch, I'll merge it to master.

Btw, is the XML parser linked to creatures stats, such as their HP definition? So that we can fix the creatures instantly dying when
fighting problem?

Best regards,

[Bertram]

Your CMakeList.txt seems to be utterly broken , and gernerating invalid Makefiles as one can see above ...

I swear it is not, at least for me. And I tested it several times with a clean cmake cache and build data.

This comes from the Cmake-gui CMakeList.txt :

In a fresh clone of my branch (sorry in advance for the time to wait & Don't forget to clone with --depth=1 to lower the data to download),
could you tell me whether the 'cmake .' invocation in a shell leads to the same problems?

[Bertram]

For everyone, Paul and me ended up on IRC.

@Paul: I'll push my code commits on development, except for the CmakeLists.txt change, since it seems you have problems with it.

As for the SAX loader, I see the point on loading entities (whatever they are) base on XML events.
To me, it will mean entities will rather have to be defined on the same level for simplification purpose.

Why? Because this way, you'd be able to do this:

{l Code}: {l Select All Code}

void ODParserXml::on_start_element(const Glib::ustring& name,
               const AttributeList& properties){
    if (name == "creature") {
        // Parse properties
        game->AddCreature(...);
    }
    else if (name == "trap") {
        // Parse properties
        game->AddTrap(...);
    }
}

This way, you could simply get rid of the visitor and the stack and have something rather simple to deal with

What do you think?

Also, we severely need to simplify those XML and make the game data structure take in account the ability to follow a campaign.
This means we need to separate the common properties from the ones linked to a specific map.

Best regards,

[Bertram]

Here is the first version of my XML map data counter proposition:
(See the comments in it)

{l Code}: {l Select All Code}

<?xml version="1.0" encoding="UTF-8"?>
<GameMap width="400" height="400">
    <!-- Different AI + Player -->
    <Seat team="1" faction="Undead" startingX="-5" startingY="10" color="0.0;0.4;0.0" />
    <Seat team="2" faction="Heroes" startingX="15" startingY="0" color="0.4;0.0;0.0" />
    <Seat team="3" faction="Villains" startingX="15" startingY="0" color="0.0;0.0;0.4" /> <!-- You -->
    <Seat team="4" faction="Constructs" startingX="15" startingY="0" color="0.4;0.4;0.0" />
    <Seat team="5" faction="Animals" startingX="15" startingY="0" color="0.4;0.1;0.4" />

    <!-- Player goals -->
    <Goal task="KillAllEnemies" options="NULL" />
    <Goal task="ProtectCreature" options="King;-1" />
    <Goal task="ClaimNTiles" options="1000" />
    <Goal task="MineNGold" options="20;-1" />
    <Goal task="ProtectDungeonTemple" options="NULL" />

    <!-- NOTE: The creature definitions should be loaded only once in another place than here. -->

    <!-- Map lights -->
    <Light pos="-4;8;3.75" Diffuse="0.9;0.8;0.6" Specular="0.1;0.1;0.1" Attenuation="50;0.012;0.32;0.0018" />
    <!-- ... -->

    <!-- Special Tiles at game start -->
    <Tile posX="-54" posY="-98" type="3" fullness="100" />
    <!-- ... -->

    <!-- Rooms present at game start -->
    <!-- NOTE: The mesh used should be resolved through the type, not be declared here. -->
    <!-- For simplification purpose, the room type should be declared for one tile at a time.
         Or, we can declare an array of tile type id, which is faster for a parser to handle. -->
    <Room type="Forge" team="1" posX="1" posY="2" />
    <!-- ... -->

    <!-- Traps already present at map start -->
    <!-- NOTE: The mesh used should be resolved through the type, not be declared here. -->
    <Trap type="Cannon" team="1" posX="-14" posY="1" />
   
</GameMap>

We could also let other xml types, describing the creatures, room, and traps definitions. The definition files would be loaded once. And the map xml loaded at game start,
and later when the corresponding map is selected.

From all I could see, I'd still get rid of the SAX way of doing it, and stick with the simpler DOM way, while getting rid of the Xincludes.

I can code the counter proposal which will be simpler to read and maintain, IMHO, and using the library the former team seemed to be decided to use: tinyXML.

Best regards,