our cmake modules have become good enough, prefer them over system ones

Backed out changeset 4cccc0520988

our cmake modules have become good enough, prefer them over system ones

write FindGHC.cmake for haskell stuff

*sigh* I always forget to pull when AFK for a while. btw. why are we still keeping binaries in here? ☹

user correct variable and slimmer test, version typos, .dll.a updated

extend time a bit. move to defrosting on team's turn. fairer than using round order I think.

comas and capitals

rework ffmpeg/libav/videorec linking and their cmake discovery

rewritten findfreepascal, moved checkstack code elsewhere

support old cmake style variables

merge

rewritten findffmpeg.cmake

replace sdl mixer/image version checking with something more robust

rewritten FindOggVorbis.cmake

fail if sparkle not found consciously

oups

rework FindSparkle.cmake

use opacity to indicate thawing. not totally happy w/ effect, but argument was made to indicate some progress towards thaw. drop "exit" since it really wasn't helpful, and not rendering tags, was more annoying than fun.

At multiple requests, change ice to be turn based. This is a more modest adjustment of existing code, that can be fairly easily flagged or reverted if desired. It also allows scripts to extend freezing, and preserves fire melting.

escape return and be more friendly with translations

add link to GPLv2 notice in about page

uniform revision info in about page

add version info to frontend output too

version variables confom to naming convention

oooooops forgot SHAREPATH is still used somewhere

lupdate-qt4/lrelease-qt4

apply variable change here as well

on linux assume datapath is always relative to cmake_install_prefix, unless you provide an absolute path; on win32/osx hardcode engine path to default installation prefixes (it gets overridden at runtime so it works even if you move the data folder)

Oops, forgot this check (fixes crash)