[Kazan]

project name is alliance
module name is pcs2

fs2_open is the SCP, and they're on different servers

[Mav]

Thanks

... Ooops, so that adress (warpcore or something like that) from the wiki is still actual? Because I got an error there, too; though a somewhat stranger one - "cvs [server aborted]: "import" requires write access to the repository"... 

[Kazan]

warpcore is not where my apps are hosted, i use sf.net

[Bobboau]

yeah PCS2 is (from a code point of view) totally unrelated to the SCP (other than the fact that it writes file for it of course.)

[Bobboau]

hey, kaz, is there any particular reason why the GL canvas doesn't accept wheel events?
probably has to do with it not accepting focus.

also I am in the proccess of adding omnipoint editing capabilities, though not drawing, in theory all you will need to do is make a function that takes an omnipoints structure and build a (set of) vertex (and posably index) buffer(s) out of it, and code to draw that (set of) buffer(s) in the GL render function. you should wait untill I commit this before working on anything remotely high level about this.

[Kazan]

i think the outermost panel has to capture wheel events... but i've never dealt with them before

[Bobboau]

I got it to work by forcing the GL window to get focus, to further complicate the issue I also captured which window had focus before hand and give it focus back on the mouse leave event. this could cause problems if the window which had focus gets deleted in the time between when the mouse enters the render window and leaves again. but I don't think that's likely, because for the user to do anything that would cause any window to be deleted they would have to leave the window.

while I was in the GL canvas I took a look at the code responcable for fixing the aspect ratio and it's the damndest thing, I though I had figured out why it wasn't working (you had it doing integer division when it should have been float) but changeing that to what it should have been didn't change anything, it's still only accepting the integer part of the passed float.

(maybe if I rebuild the solution it's just magically fix itm [edit]nope, to see what I meen, load a model, if the render window get's taller than it is wide the model will disapear, if it get's two or three times as wide than it is tall it will fix it's self but only for 2.0 or 3.0 (ect) aspect ratios[/edit])

[Bobboau]

ok, I just committed today's work, right now it would probly be a good time to work on omnipoint rendering. my recommendation would be to add code to wxGL_PMFCanvas::refresh_omni_points (called when ever something on the outside made a modification to the internal omnipoints) that makes one or more vertex/index buffers from wxGL_PMFCanvas::omni, and add code to wxGL_PMFCanvas::Render to draw these buffers if they exist. omni is a struct which represents a 2d array of omnipoints, it in addition to the array also has render (and data type) flags, and colors for unselected points, points in the array which is selected, and the point in the selected array which is selected. hopefully different colors can be done with the same vertex buffer. it might be worth it to have the selected point be drawn on it's own so when it's moved nothing else has to be constantly rebuilt (internal modification of the omnipoints is not taken into account with regards to the refresh_omni_points method). I moved all omnipoint related stuff into it's own header (including a color class that could probably be better off somewhere else). omnipoint editing should work right now unfortunately you can't see what you are doing, don't worry too much about making it look pretty at this point, once I have some basic code in place I'll see about making a geosphere based indicator.

[Kazan]

feel free to make those changes yourself - openGL is easy and straightforward

[Bobboau]

well if this was an API I knew I could do it in an hour or two, I don't realy want to learn a whole new graphics API just to draw some spheres. I'm suposed to be the UI programmer not graphics, and I've gotten everything but the low level drawing code done.

[Getter Robo G]

Heh, you guys deserve your own Project Badge for this: (PCS2).

 

[Bobboau]

tell you what, I just doodled some crap code that seems like it can make a vector of vector3ds (points) and one of int triples (representing triangles) and a third vector which tells you were to start looking in the triangle vector for a geodesic sphere of a given resolution. I also have some very straightforward formulas for telling you how many points and triangles should be in a sphere of a given resolution. I was just curious as to how complex a process it would be, hence the doodle code (nothing actualy directly useful), but the basic idea behind it might be useful.
if you could just give me an entry point, a graphic vert struct that has position and color, and a function that would take (1) an kaz_vector of those and make GL's equivalent to a vertex buffer and (2) a vector of shorts and make an index buffer, and then set those to be drawn I'll do all the procedural geometry stuff. all I need you to do, is the lowest of the low level API stuff.

[Bobboau]

well I'm working on drag and drop now, it seems that icons for all items in the tree are a requirement of this, so I've had to make a bunch of icons, and as a result I have a nifty batch file for  converting normal editable graphics into XPM

Code: [Select]

@echo off

:LOOP
if   "%1"==""  goto  DONE
set str=%1

convert %str:~0,-4%.%str:~-3% %str:~0,-4%.xpm
shift

goto LOOP
:DONE

this batch file requires you have Image Magick installed, you just drag one or more files onto it and it'll convert all of them.

also for things like the insignia, I have thought for a long time that the way to go with that would be to have a non-pof data chunk representing a few boxes that would define were the insignia should be, and then to have the editor build the geometry when saveing as POF, given that PCS has a native non-pof working file format, I think this would work even better, however there is no way to reconstruct the insg boxes from an exsisting insg chunk, so importing of insg would be lost, unless we did something tricky.

[Bobboau]

for everyone else, this is how drag and drop is going to work.
there are three levels dragability, the first level is undragable items, these will be the chunks, draging a chunk makes no sense cause there is no were for you to go with it. the second and probably most intuitive will be abstract data, abstract data is any collection of primitive data which exists in arrays, this will be things like your weapon points and your paths and stuff like that. abstract data will generally have the most drag and drop flexibility, as it will use DND primaraly as a means to allow hierarchy modification, for example moving (or copying) a (posably more than one) thruster point from one bank to another. abstract data will only be dragable if it makes sence for it to be, if there is no were to go with it you won't be able to drag it, for example you will be able to move the verts of a path, but not the paths themselves (reordering is a seperate operation from draging) because there is only one array with paths in it, this means some chunks will have nod draggable items at all, like specal points and eyes. the final catagory will be primitive data, this will be your ints and strings, the only thing you will be able to do with primitive data is overwriteing one primitive data's contents with another. for example, selecting the position vector of a thruster and draging it onto the position vector of a special point will move the specal point to the location of the thruster. this final catagory is the lowest priority and might not ever get done, but I't what I have in mind.

[Bobboau]

ok, you people better apreciate how screwed I'm going to be in five hours cause I stayed up till 3am when I have to wake up for work at 6 working on this.

new build
you now have all, or most, of the submodel editing abilities you realy realy wanted, you can rearainge the hierarchy (only for submodels) and if you hold down control you can move them on the screen, cause I hooked them into the omnipoint editing system. so you can prety much do anything sort of polygon/uv editing or rotateing submodels. yes, your welcome... there will probly be repricusions in terms of bounding boxes needing to be recalculated or something, but that is for another time.

and there seems to be a slight bug in my omnipoint code in that it will assign the value of the last thing you were moveing to the thing you just started moveing, consitering this is the first visual feed back I've got with the system it's not suprizeing, I'll fix it tomorow.

[Mav]

ok, you people better apreciate how screwed I'm going to be in five hours cause I stayed up till 3am when I have to wake up for work at 6 working on this.

    ...

... so you can prety much do anything sort of polygon/uv editing or rotateing submodels...

Take it as appreciated... wow! 

And Getter Robo G is right - you should get a badge for this .

[Bobboau]

eh, that should have been short not sort. I was half asleep when I wrote that post.
so, you can not do that yet. but you will eventually.

anyway, is nobody reading this? I would have expected more of a response by now, that's the sort of feature people have been begging for for years.

[RazorsKiss]

It's still afternoon, in the US.

You'll get some more later today, I'm sure.  It looks very nice thus far.

[Spuddy2000]

Yes I am.
Have downloaded latest version.
Works out of the box no textures though!!
Yes I have paths for textures in the .ini file.
Good to have all the options of PCS!!
I can't see any selection indicators on my main model screen when I select something though!!
Is this intended or a problem with me??
I have Windows 2000 and Inspiron 8000 laptop with 500Meg RAM. ATI graphics chipset.

PCS2.....Its a great modelling tool........ upgraded to be even cooler so bozos like me can use it!!!Keeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeep Goiiiiiiiiiiiiiiiiiiiiiiiing

 

[Bobboau]

if by selection indicators you mean something in the 3d window to show the various points of the POF data, then that has not been implemented, I don't know Open GL and kaz seems to be refusing to write the last bit of code for that for some odd reason, guess he's busy with something else, but it should be working in a build or two.