greetings.
quixote
project a bit.
First thing I am going to work on is getting something to work in
linux (either natively or through wine).
Looks like somewhere in the jump from 0.8 to 0.9 it doesn't like linux
anymore :(
I've also got a Macintosh at home I can set up and get that running.
Anything else I should be aware of?
Roger Gammans
> First thing I am going to work on is getting something to work in
> linux (either natively or through wine).
> Looks like somewhere in the jump from 0.8 to 0.9 it doesn't like linux
> anymore :(
I'll put them up in my hg repo soon.
There where two issues, 1) The deckbuildier uses the Native
win32clipboard module not the wx.Clipboard
object. I haven't change this yet but it is easy to comment out the
function which rely on win32clipboard
for the time being.
2) There seems to be an issues with wxgtk in that LIST_ITEM_BEGIN_{R,}
DRAG events don't fill in the Data
field from the selected Item so, I had to change the way the Dragged
cards Id is found. This is a 6 line patch.
The only other thin was I had to make sure I had the python opengl
bindings installed.
ro...@computer-surgery.co.uk
On Jul 2, 10:33 am, Roger Gammans <roger.gamm...@gmail.com> wrote:
> On Jul 1, 5:20 pm, quixote <quix...@toysmakeuspowerful.com> wrote:
> > Looks like somewhere in the jump from 0.8 to 0.9 it doesn't like linux
> > anymore :(
(SEGVs) while
initilaising the GLCanvas object.
> I've just done some patches that make it run fine on Linux for me.
> I'll put them up in my hg repo soon.
First is at http://hg.backslashat.org/repos/eopk-svn/ , whicis just
a
import of the svn tree.
Then based of that ismy changes in this repo:-
http://hg.backslashat.org/repos/eopk-linux/
> There where two issues, 1) The deckbuildier uses the Native
> win32clipboard module not the wx.Clipboard
> object.
http://hg.backslashat.org/repos/eopk-linux/rev/99ef26405cd6
> 2) There seems to be an issues with wxgtk in that LIST_ITEM_BEGIN_{R,}
> DRAG events don't fill in the Data
--
Roger.
Don Eisele
>(SEGVs) while
>initilaising the GLCanvas object.
That's what's been happening on all 3 of my ubuntu boxes.
--
Don Quixote -- qui...@toysmakeuspowerful.com
Tatsumaki Aikido: http://www.k-state.edu/aikido
Woodworking Gallery: http://toysmakeuspowerful.com/woodenplunder/gallery
CCG Card Singles: http://imperialblock.com
L5R Trades: http://trade.mahasamatman.com/L5R/list_show.pl?user=eiseled
TengaiCON game convention: http://tengaicon.com
Roger Gammans
>
> >Interestingly I haven't got it to run on my laptop yet. It fails
> >(SEGVs) while
> >initilaising the GLCanvas object.
>
> That's what's been happening on all 3 of my ubuntu boxes.
Interesting. Are they 32 or 64 bit.
Are all the relevant pkgs you've installed Unbuntu
signed?
The biggest diference between my laptop and my office desktop
is that my laptop is amd64, and my office machine is still i386.
(In terms of the output of `dpkg --print-architecture` that is)
TTFN
--
Roger. Home| http://www.sandman.uklinux.net/
Master of Peng Shui. (Ancient oriental art of Penguin Arranging)
Work|Independent Sys Consultant | http://www.computer-surgery.co.uk/
New key Fpr: 72AF 0ACC 9A53 E59F B1B6 DC14 1983 A13E 5C3D 3CEB
Roger Gammans
> Interestingly I haven't got it to run on my laptop yet. It fails
> (SEGVs) while
> initilaising the GLCanvas object.
Ok. I've traced this a little over the weekend and it
is occurring because the dispatch table in mesa isn't setup
(see http://www.mesa3d.org/dispatch.html ).
I'm guessing this means there is a missing OpenGL initialisation
call but I don't really know the openGL api to know what Call is
supposed to set up the dispatch table.
As far as I can tell it should be wxGLCanvas::SetCurrent() but we are
AFAICT calling this. (And adding and extra precautionary one doesn't
seem to help).
Does this help anyone else?
--
Roger Gammans <ro...@computer-surgery.co.uk>
Roger Gammans
> On Fri, 2009-07-03 at 02:07 -0700, ro...@computer-surgery.co.uk wrote:
>
> > Interestingly I haven't got it to run on my laptop yet. It fails
> > (SEGVs) while
> > initilaising the GLCanvas object.
There is an update on my mercurial repo at
(http://hg.backslashat.org/repos/eopk-linux )
This might help - I and I would certianly be interested in
reports of success or otherwise with it.
It fixes a blank card backs problem I see with the i386 build
of mesa.
I've also report this as a bug to the mesa team at debian, as
bug #536106 with a proposed fix.
As is possible as I mentioned ealier this is a entirely an openGL
issue.
Don Eisele
glMatrixMode(GL_PROJECTION)
32bit ubuntu 9.04
That's on my work machine... I'll confirm if it's the same on the
two laptops later.
On Tue, Jul 07, 2009 at 07:03:28PM +0100, Roger Gammans did say:
Don Eisele
to wad through, so that's where I started)
On Wed, Jul 08, 2009 at 02:38:02PM -0500, Don Eisele did say:
Roger Gammans
> (This is also the deck editor.. it was smaller and easier
> to wad through, so that's where I started)
Hm. I'm pretty sure the deck editor worked for me straightaway except
the clipboard.
And I've never had any 32bit hangs. I lost the card
backs which - and those glGenTextures calls , was where the main routine
was crashing for me. Which I why I moved them to ensure they happen
after InitGL( ).
You don't have some sort of animated opening for windows setup do you -
If your eopk is rx-ing resize events before the first paint odd stuff
may happen.
I suspect most of the problem is when gl calls occur before a Paint
event which is he first time a x-platform application can be sure to
have windows DC (needed for selecting GL context)
I hesitate to change my mesa back to the distribution one now I've got
eopk working by patching mesa.
--
Roger Gammans <ro...@computer-surgery.co.uk>
Roger Gammans
> Looks like I'm getting my SEGFAULT in preview.py on this:
> glMatrixMode(GL_PROJECTION)
>
look at the values of _gl_DispatchTSD and _glapi_Dispatch when the
segfault occurs.
You'll need to install the libgl1-mesa*-dbg libraries though to get
access to the symbols .
TTFN
--
Roger Gammans <ro...@computer-surgery.co.uk>