[Mav]

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.

Oops... But still it's great . And as long as you are planning to implement those, well...
Though I'll most likely stick to tS for these things, anyway .  (prepares to defend against tS-rants...  )
It would be nice to have the option to quickly correct small errors (that you've recognised too late) this way, though.

[Water]

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.

Seems to work ok, but closes tree after each move.

Unrelated:
Your DevIL ver 1.6.1
Kazan uses ver 1.6.6
New one crashes if devil  *.dll files not 1.6.1

Saving large pofs seems to work now. 
Textures don't show on vLarge objects. (black view) Probable auto zoom guesses wrong

[Bobboau]

yeah the tree closing is sort of annoying, it's weird that it's doing that I would expect either everything to close or nothing, I'll have to look into that.

hmm and here I thought I was using the most up to date DevIL, maybe I just grabbed the wrong DLLs.

[WMCoolmon]

I'm just playing with PCS2 in an empty state, but it seems like
A) Every time I add a gun point, my missile point will disappear (And vice versa)
B) The tree starts out completely collapsed and way too small

[Nuke]

im still getting that uv bug.

[Bobboau]

hmmm... yes, weapon points do seem to have a bug there, I hate haveing to deal with weapon points, Kaz made a bad design choice on how he handled them... rather than two separate lists there is one list and you have to differentiate everything based on it's type (an int in the bank structure), I must have simply set the list in the model to a sublist containing only the items in the control, I'm suprized no one else noticed that bug yet.

not sure what you mean in part B, there should be about a dozen root nodes each one colapsed (cause there is nothing in them, except for the weapon points which is divided), is this not how it is for you? what do you mean by 'way too small', you mean how thin it is? there was some sort of odd bug with the resizing frame I haven't been able to resolve were that frame will not get any smaller than what I initially set it to be.



UV bug?
loading textures upside down?
if that's the case I bet it's DevIL, it does that for some things, for no good reason.
what is the texture format? (I know it load TGAs upside down)
were has Kaz been? some of this is stuff he needs to look into.



ok everyone, I want you to download tonight's build and I want you to load up a model and I want you to go to the path's section and I want you to look towards the bottom and tell me if you see anything interesting.

[Nuke]

im a little drunk right now so im probibly not too good at debugging things. but my texture format is usually dxt 1 or 5, and in the case of the ships i tested, both formats were used. seems to be a case of texture inversion.

[Water]

hmm and here I thought I was using the most up to date DevIL, maybe I just grabbed the wrong DLLs.

Sorry my mistake - Kazens dlls work ok with your one. I had a DevIL rc2 in there by mistake (this was causing the crash)

[WMCoolmon]

Attached is how PCS2 looks when I first start it up. (New build). Note how the tree is almost completely nonvisible, you have to drag it out. It's a minor thing but is probably easy to fix.

I don't see what use there would be in making the tree get any smaller, if you want it to disappear entirely you'd be better off relying on a menu item or a button, which you could assign a hotkey to. Hiding the tree and settings pane along the right might not be a bad idea, it'd help people with precision placements of points using the graphical interface, or just taking pretty pictures of ships.

[attachment deleted by admin]

[Nuke]

well id put a little tab on it that you could click to expand the tree withut the need of a drag and drop event. the tree would allocate as much space as it needs to display everything without scrolling.

[Kazan]

hmmm... yes, weapon points do seem to have a bug there, I hate haveing to deal with weapon points, Kaz made a bad design choice on how he handled them...

i didn't want to putz with code duplication or function redirection (IE look at how POFHandler.h handles GPNT/MPNT and TGUN/TMIS) so I made them one list of "hardpoints" with a type flag

[edit]
some notes

A) Memory leaks increased significantly when i CVS-updated this morning
B) Scroll wheel zooming needs to be faster
C) Please tell me what features you have implemented from the feature request thread so I can mark them "in testing"
D) Please use the PCS2 Test Builds Thread for posting builds

[Bobboau]

well all I know is with that setup weapon points, which should be one of the simplest bit's of data has been practically the bane of my existence, it's probably too late to change it now though.

A)when was the last time you had updated? I've seen a bunch of memory leaks for a while (usually like <4 bytes) I have been under the impression that they are coming from wxWinows. but I have been on a lookout for anything that could be causing it. but I don't use the new operator much (if at all) except for when making new windows, which wxWindows then tells me it'll take care of.

B) I agree, I'm intending to focus on the 3d controls a lot more soon.

C) all data is editable now, more or less (maybe one or two things which I overlooked, or things which get ignored).

I just wrote some auto-path code, it's not 100% done, but the feature most people are going to use is done. I don't have a thing for setting a given path automatically, only genorateing a new set of paths, which nesisaraly erases existing paths (I exempt docking bay paths)

I can probably write a gunpoint convergence calculator after I take a nap (which is what I have scheduled next on my itenerary) that should be ridiculously simple.

C) had forgotten about that.

[Kazan]

from what i've read i think they're false positives from CrtDebug... it's the memory manager fragmenting.. it's not really leaked, it's just unallocated memory that hasn't bee reclaimed yet

[Kazan]

oh bob... they should be able to store comments in PINF so make it editable

[edit]
important note for geometry editing.  if you edit the geometry of a subobject you MUST call PCS_Model::SetObjectChanged(unsigned int idx) with that subobjects number otherwise the BSP cache, if present, will still be used when writing to a POF

(this is a new function that will be in CVS soon)

also making it so PMFs write their BSP cache to disk, and when BSP cache is present and a SOBJ is changed when it's saved to POF again it'll cache the new BSP.

this ups the PMF file version to 102 (not that that matters to you)

also found a bug in PMF writing - forgot to update the output version number to 101 when I implemented those changes.. so any version 101 PMFs written before now would read in corrupt (because the reader sees version 100 so skips reading some data that is present in 101... gettting offsets all fusterclucked)

[Bobboau]

PINF editing should be so trivial it'll take me less than ten minutes, if that.
in fact, lets time it.

[Bobboau]

alright, more or less done, I didn't realise till the last second that PINF was a list of strings, I thought it was just one so I'm going to have to go back and make a multi-line string list, in stead of just a multi-line string.

hmmm. took a little longer than I was expecting, 15 minutes.

[Kazan]

i assume the current tree graphics are place holders?

[edit]
oh... and the mouse scroll... way too fast now.. model went into oblivion and won't come back.. we need a reset for that

[Bobboau]

they are placeholders that are good enough if we don't think of anything else.

working on the scroll wheel...

[Bobboau]

anyway, I'd realy like the 3d stuff, so...

ok, if I'm gona draw something, it would be like

Code: [Select]

glBegin(GL_POLYGON);

for(each vert in the poly list){
glTexCoord2f(u,v);
glNormal3fv((GLfloat *)/*pointer to a vector*/);
glVertex3fv((GLfloat *)/*pointer to a vector*/);
}
glEnd();
???

[Bobboau]

linking error?... LINKING ERROR!?!?! well, gotdamnit you knew good and well that was suposed to be inline'ed! MSVC quit screwing with me!!!!