[Freeciv-Dev] Re: (PR#12111) Xaw3d client startup problem(png problem)

["Vijay Kiran Kamuju" <infyquest@xxxxxxxxx>]

<URL: http://bugs.freeciv.org/Ticket/Display.html?id=12111 >

currently the cursor bug is in the way to test this one.

unless that is sorted out, my civclient does not yet start.

i can check whether the fix is working or not

bye,
Vijay

On Sat, 5 Mar 2005 03:05:27 -0800, Egor Vyscrebentsov <evyscr@xxxxxxxxx> wrote:
> 
> <URL: http://bugs.freeciv.org/Ticket/Display.html?id=12111 >
> 
> On Fri, 4 Mar 2005 09:42:16 -0800
> Jason Short <jdorje@xxxxxxxxxxxxxxxxxxxxx> wrote:
> 
> JS>Vijay Kiran Kamuju wrote:
> JS>> i am trying to make necessary changes to graphics.c,
> JS>> by writing a new function generate_palette a modified version of entire 
> pngquant
> JS>> its quite a bit blown up, not completed, non working,etc
> JS>>
> JS>> I am also trying to create a separete file for gen_pal.[ch] a modified
> JS>> version of pngquant.[ch],  so that al the functions of orig pngquant
> JS>> are encapsulated in one function,gen_palette
> JS>>
> JS>> i thing it would better to call the function as convpng_rgba2rgbap()
> JS>> or pngpalette_gen rather than gen_palette
> JS>
> JS>I'd suggest instead changing the code to work natively with non-paletted
> JS>PNGs.
> JS>
> JS>1.  Find some GPLd code that loads a PNG image into a simple graphics
> JS>format (4-channels, 8-bits per channel, rowstride, etc.).  I know Cairo
> JS>has such a utility function but it may not be GPL.
> JS>
> JS>2.  Once the image is loaded into this form, you can go over ALL the
> JS>colors to allocate them.  Possibly this should only be done on paletted
> JS>displays (there's probably a way to query this in X; I know GTK has it
> JS>so you can probably just look in the GTK code).
> JS>
> JS>3.  Eventually masks with alpha can be supported as well.
> 
> Well, i have part of graphics.c code which load RGBA images, but i haven't 
> get it work full.
> From last changes (grid?) i even can't get xaw client started. And don't know 
> why.
> And i don't understand now alpha masking...
> 
> But anyway, attached is a first attempt of patch (with codestyle and other 
> errors).
> 
> Thanks, evyscr.
> 
> 
>