Win64 build , please test it , cause i had to use some hacks to get mingw64 work with cmake

Here's my fork (dunno where the original went) of DasBruce's wx frontend. Maybe you can coordinate/re-use some stuff of his when you actually tackle a GUI.

edit: will prolly have a buntu package for this sometime later today

If you do it it's ok , cause i didn't like wx last time i used it, otherwise i'll do it with Qt , the map compiler can be used as library too , if you want i can add function to pass raw data as image directly

I meant a package for your mapconv. The inital debianization is done and it's currently waiting to be built on launchpad: https://launchpad.net/~koshi/+archive/ppa/+packages
If someone wants to try it right now, here's a amd64 build from my desktop box: http://filebin.springlobby.info/mapconv ... _amd64.deb

I wouldn't start a new GUI project with wx either and would prefer Qt btw.

Maybe spring could be hacked to support map creation directly

Be careful ,the "decompiler" target has a too generig name , maybe rename it to smfdecompiler

Right, will do. Mind if I lowercase the other binary target?

so even with compression off you match 100% equal tiles. Ok what about compression on, you say no artifacts but at high compression how do you avoid that?

If you lowercase the name , decompiler generated makefiles won't work

I've added a nonlinear matching , that if a single pixel is more than 30 different ( noticeable by looking at it ) , it will count like 50 pixels different , so it is unlikely to be reused , unless you use more than 0.8 as th value

if you use -o ./somedir/ map texture will be broken.

By the way thanks, launching mapconv under wine was pain :)

Forgot to mention that I've made this available via the spring ppa: https://launchpad.net/~spring/+archive/ppa
Maybe you could start tagging commits that you deem notable advances? I'd then update the package for those.

wiki page: http://springrts.com/wiki/MapConvNG

can someone with the knowledge write a howto for mapconvng + deprecate all mapconv howtos?

Hmm, SpringMapConvNG.exe keeps crashing for me. To see if it was a problem with my map, I tried using decompiler.exe on DesertTriad, and then running the make.bat it spits out. It gets to "Compress done, ratio 99.4446" and then will either crash or seemingly quit normally. Either way, all that gets written is an empty (0 bytes) DesertTriad.smt and no DesertTriad.smf
However, it did work once with my own map! I just can't seem reproduce that. I'm running Windows XP.

So nobody else has this problem? I thought maybe it was because I'm running XP, but i just tried it on a computer running Vista and it does the same thing. The strangest thing is that it crashes 90% of the time, but every once in a while, it works, and if I use the exact same command line options again, it will crash again. The one thing that seems almost consistent is that if I specify a minimap, it crashes after loading the minimap, and never gets to compressing tiles. But then, every once in a while, the minimap works, too. Am I in some sort of parallel universe?

From Windows Event Viewier:

So I have a Texture Map, Metal Map and Height Map that compiled fine on behe's mapconv, but when I compile it on Linux with mapconvng it doesnt complain and when I open it in Spring I just get the flat Texture Map rotated 180 deg and submersed under the water. Any Idea what is going on? Thanks in advance.

10 secs. So what were the other guys doing before, a little folding@home maybe..

I'm confused...

the win32 build crashes for me, too.

can someone please rebuild the current source?


then it crashes.

Output looks weird, maybe my commit fixed that already. (https://github.com/tizbac/SpringMapConv ... 162#L0R700 )

to reproduce: just decompile / compile http://api.springfiles.com/?springname=DesertTriad on windows (with wine it didn't crash)

gdb output: http://pastebin.com/WNUjqp03

i wasted more time on trying to fix the crash on win32:

https://github.com/tizbac/SpringMapConvNG/issues/6

but didn't find a proper solution.... only commenting out free() fixed it for me. any ideas whats wrong? (see the bug report)