Remove unused global variables

*sigh*

merge

A bit of refactoring

Use HW_TEAMCOLOR_ARRAY from frontlib

Invert S3D_DISABLED to more consistent USE_S3D_RENDERING

Merge

GCI: use player's name for the default team task: http://www.google-melange.com/gci/task/view/google/gci2012/8007227 note: these bugs were NOT introduced by this: - double confirmation on team removal - team addition fails

Fixed flib_weaponset_create creating an empty weaponset (for real this time :P)

unbreak linking pending some fix or other by koda. this is the same thing video recording still does.

Some fixes

Added default weaponset to frontlib hwconsts Now the weaponset create function can actually do what it promises and create a default set instead of an empty one :)

merge

No more crashes when running quick game

GCI2012: Empty list help text (issue #375)

Try to setup weapon sets for quick game

Remove limit on hedgehogs number

Merge from default to have recent fixes

koda, koda... could have checked your claims

link physfs on ios, add its library and (most importantly) implement automatic dependencies resolution

Updated Slovak translation

add ios project to physfs

merge

INSTALL and README updates

physfs: no carbon depency on osx, add cdecl attributes on extra functions

Updated Slovak translation

Load sounds with the help of physfsrwops library

merge

Refactor identifiers names

GCI2012: Check for Clang Version in CMake

Refactor Clang module.

Implement QAbstractFileEngine::FastReadLineExtension

Move version check of Clang to CMakeLists.

Use alternative way of checking for file end. Also refactor this code a bit.

Move code for Clang to separate CMake module.

Move code for Clang to separate CMake module.

Bring ansistrings back

Oops, temporary objects involved

ok at least pas2c now parses allnots.......

show console on win when debugging to see frontend messages

Convert HWTeam to be a wrapper to flib_team

update var name

Oops, sorry, libphysfs.dll

- frontlib loads team config via physfs - some more work on frontlib port

throw in some 'nots' trying to restore pas2c functionality

- Fix saving .ini files - Fix frontend effects value loading - Fix some places working with files not via physfs

Remove debug messages

updated branch

revisit the debugfile section

Fix some frontlib bugs

moar cmake cleanup, use message(WARNING ...) where appropriate

Try to use map preview facilities: - HWGame and HWRecorder are completely broken - HWMap should work, but doesn't

Follow IANA suggestions better

cleanup revision section, make hg launch tolerant to config errors, drop deprecated exec_prog for desktop configuration in favour of execute_process

Bad unC0Rr

Bad koda!

little cmake cleanup, search installed modules first, then our own

how this happened?

Take care of parameter passed to FileEngine::fileName. Fixes problem with weird paths on windows.

Some fixes to build process

added sdl_net headers and def lib for windows, fix frontlib compilation on windows

lotsa hackery to get frontend somehow link libengine...

frontlib paths

fix physfs on osx, now all platforms should be happy

Oops

Start port to frontlib

Fix build

Fix SDL library name

find lua.h ...

thou shall not use system headers for crossplatformness

physfs compilation on windows * external calls must always be named * physfs has to be compiled as dll * there shouln't be external variables, implemented a function that sets buffer * physfs extras is now integrated in main physfs * removed physfs extras

use lua headers correctly

Fix typo

Okay, remove previous request only if it has same parent as this one. Fixes the last note of previous commit (which was nearly impossible to hit, but whatever, just cleaning implementation)

- Only allow one engine instance running at the moment - When adding new map preview call to the queue, check if previous map preview request has started, and if no, remove it from queue. Could be harmful if you manage to request previews for two widgets in less time than needed to generate one.

win build changes, fpc 2.6, CMAKE_PREFIX_PATH, glut header

Link some libraries needed for physfs on windows (there are still 5 symbols which couldn't be found - to be resolved)

Merge

fix deprecation warnings for libav sources

meh

linking phsyfs on osx

allow the engine to use physfs on android too

Merge from 0.9.18

Convert reading hedgehogs limit from map config

Download .zip too, and store them to .hwp

merge... need to remember to update first

Merge physfslayer branch into default. Have fun!

Close this branch

- Download .hwp files into Data folder and mount them immediately (yes, could cause problems in case of overlapped contents) - QuaZIP no more. RIP.

oops

oops

Mount .hwt files found in Data folder

oft-requested, should make the shoppa guys happy, probably, but, knowing them, I'm sure someone will complain

link main on osx

Get rid of DataManager::findFileForRead

Backed out changeset 5d00c28059e3 - enable the feedback button again

cleanup headers, remove -I . from rtl lib to save a few warnings

Use (undocumented) Mix_LoadMUS_RW to load music

mutatis mutandis

some CMake cleanup

We need custom script loading function in lua now

up android target api (but this commit is actually just to test buildbot: desync

fix build

Keeping bruteforce

another tweak tomake --stats-only work on headless servers

and install it too

Oops, didn't realize FindLua.cmake is our own module

Don't rely on hedgehog facing direction, check intentions instead. Should fix excessive turns around and weird jumps to water

comments and runtime list of files to convert and compile

Implement reader in C ffs

Pretend AI knows how to use drill rocket (actually has no idea of drilling)

finally hwengine in c links without errors

Script loading via physfs which doesn't work: - Some scripts just refuse to load for no obvious reason - Some scripts load fine when you set BUFSIZE to a large value, like 64k, but don't load with small values Error messages in both cases don't seem to have relation to lua file contents

Call fallers less often. Should hopefully still be about as effective at the intended purpose. Should help loads quite a bit.

cmake linking for engine in c

Engine loads fine with basic config

make stats-only mode work headless. also skip a few things to save time/memory.

merge xymeng's gsoc engine with a few updates (and further checks on symbol definitions)

Get rid of Pathz and UserPathz

oops