Changed frontlib to use the existing ini file formats of the QtFrontend

Couple of missed sections from last time. Shield is still a bit odd

Modified some crate positions, added sound to teleports, fixed loc errors...

frontlib: Rewrote the ini helper code, finished ini reading/writing support for all relevant file types

reorganize options.inc a little

Added localization and changed primary team color.

Cleanup, refactoring and generally more development in the frontlib

uniform NOPNG and NOSERVER cmake symbols

Added Mission 4: United We Stand

Merge

sed 's/\t/ /g' CMakeLists.txt

Set the choice constants to the correct values (1-3 instead of 0-2).

Ensure that Clan ammo always takes precedence over Per Hog Ammo in the flags. It'd be best for the user sake if these buttons acted more like radio buttons.

Removed some debugging lines I forgot to remove.

disable this debug too

Added mission 3: The Journey Back

Remove debug pixel draws, skip tracing on own shoved hogs. Should speed things up allowing more evaluations of stuff.

Fixed some bugs with Mission 2 and Animate.lua.

TestBaseballBat

Fixed a bug in Mission 2, where hidden hogs appeared before getting deleted on losing the game.

- Adjust some consts so whip and firepunch work perfectly - Temporarily make AI draw expected traceshove trajectory while I'm working on baseball bat

Changed color of the "Natives" team to make it more visible and reduced difficulty of the "easy" challenge in the first mission.

Ignore all objects for fall tracing with shove to avoid considering checkins as obstacles. many objects will get knocked by the kick anyway, so end result should be pretty good. Oh, and ditch the sniper rifle doubling.

Fixed a bug where, after finishing a campaign mission, the user needed to reselect the team in the frontend in order to see the next mission.

Double sniper rifle damage (AI seems unaware of second shot, when playing). Spread out flakes along X when looping and add a little Y variation. This avoids noticeable patterns. Especially on the rarer close layer or dense flakes like rain (for Y)

Added the helper script I forgot to add last time

- Fix damage calculation in TestSniperRifle, aim a bit lower to compensate initial angle shift in sniper rifle. As a result, AI seems to never fail sniper rifle shots. - Temporarily make AI tell its expectations in chat

Copied the first two campaign missions and the helper script into share/hedgewars/Data/

restore fpc 2.4.2 compatibility

The first campaign commit with a lot of changes... branch 'default' added QTfrontend/campaign.cpp added QTfrontend/campaign.h changed QTfrontend/CMakeLists.txt changed QTfrontend/game.cpp changed QTfrontend/game.h changed QTfrontend/hwform.cpp changed QTfrontend/hwform.h changed QTfrontend/ui/page/pagecampaign.cpp changed QTfrontend/ui/page/pagecampaign.h changed QTfrontend/ui/page/pagesingleplayer.cpp changed hedgewars/uCommandHandlers.pas changed hedgewars/uIO.pas changed hedgewars/uScript.pas changed hedgewars/uVariables.pas

New bright and shiny TestFirepunch which actually works

fix shift key by disabling the key modifiers for now

1. Add IFDEFs for video recording 2. Options for video recording were hardcoded in engine, now they are hardcoded in frontend and passed to engine thru command line (later it will be possible to change them in frontend)

- TestSniperRifle ftw (still needs some consts adjustments, because AI seems to love sniper rifle too much) - Also optimize TestDesertEagle a bit - Also some formatting changes

merge

Check for gstHHDriven instead of damage check, so ballgun stops when turn ends in multiattack mode

merge

ignore keymodifers other than ctrl/alt/shift

removed some writelns

first attempt at implementing support for keys with modifiers to use it, the keybinding have to be of the form mod:<modkey>:..:<modkey> <function> <keycode> for instance in an ini file change findhh=h to findhh=mod:lshift h

Make it AI survival almost definitely impossible by adding a 3rd pass to FindGear. 1st pass normal, 2nd pass allow close to objects, 3rd pass, allow overlapping objects.

Not needed here

Removed leftover FFP code: We should now be GL2 GLSL1.30+ forward compatible. Added a seperate shader for the gradient used for water drawing. Disabled 3D mode for GL2 codepath as this still has some FFP left.

- Allow camera movement while current hedgehog is falling - Some tiny changes

Merge

Android: first attempt at using the density call from java

only reset those keys which actually change their bindings

tidy flags

Small improvements to the frontend lib for better debugging

Welp. We have a bit of a problem here. Reverting the move of nextturn. But this causes a desync in scripting unless we can find a better fix.

here it is

Implemented game launching API for the frontlib. It is still buggy though, and not all game settings can be conveniently created/modified yet.

Don't make unnecessary calls to TestWhip, TestFirePunch, TestBaseballBat and TestHammer functions as they have no need to take Targ parameter into account and thus may only be called once per position

Implemented public API for letting the engine render maps

Make watching AI v AI on ai survival a bit more entertaining

Added ini reading/writing for game schemes to the frontend lib

Workaround for bug #144. This workaround had occurred to me a while ago, but wasn't sure if placing them unfairly was better than not placing them at all. Argument for not placing at all is people should probably abort the game when they notice it. Argument for placing unfairly is people can still abort, and if we really wanted them to abort, we should probably just have halted launch if all hogs failed to spawn. This way at least play can continue.

Frontlib: Work on the callback mechanisms for IPC Work commit; It compiles and runs but will need a major overhaul tomorrow.

After experimenting with a long running average at maxed out FPS and a variety of map sizes, 128 seems to actually be a good size to use if only drawing bits of world with stuff in it. 64 actually did even better in some situations, but significantly worse in others (lots of land, zoomed out).

frontlib refactoring socket.h now completely wraps all the lowlevel neworking, so it would be easy to switch away from SDL_net if needed. Also reduced global state by making an IPC connection an object-like thing.

Only create textures for non-empty LandPixel chunks. This should save a fair amount of memory, especially on smaller maps, and eliminate a number of draws

Merge

warp sound when AI survival hog respawns. attempt at a bit of a crate spawn animation (moar sparkles and a quick fadein)

Android: added a callback to java to determine dpi/dip how much we should scale the ui

Merge

add onGameTick20 to basic training, extend laser sight out way more (it was visible at top when completely zoomed out), move call of new turn to after AfterSwitchHedgehog to avoid lua issues in onNewTurn - if this causes problems, lua can do delayed actions in onGameTick

Some brainfucking code which greatly reduces number of TestCollision* calls in hedgehog walk routine. Especially helpful to AI optimization. Also fixes some edge cases.

merge

Demo recording for the frontend library

Name the flags for fall tracking and indicating whether the explosion erases terrain.

Moved frontlib into project_files

merge

Further work on the frontend library, restructuring, ...

merge

Add a GetTag method so that SetTag doesn't feel lonely.

Started work on the frontend networking library

Improve TestWhip

minor tweaks for options page; should improve readability at small window size... a bit...

Removed dead/deprecated code and comments

pas2c stuff

This should make drawn maps rendering even more faster

Merge

merge

Render drawn maps 10-500 times faster

Oops, forgot to set 'erasing' flag on decoding

Fix a bug with wrong state of teams list when color isn't changed manually

Remove a couple more timers, should mean alpha matches old pulse time now

also change specialists

Set air attack to active by default, make time look a bit nicer, avoid plane buzz activating.

change the array which represents the keyboard state to be a boolean array rather than byte array

Don't allow for multiple key up(or down) events, it will ignore the excess events

Also change Highlander and Racer

moved generic matrix code to uMatrix.pas updated shader and client logic to use a mat4 rather than ftransform or gl_ModelViewProjectionMatrix.

merge

merge

Add onGameTick20 to try and encourage lua to be a bit more efficient. Rewrite Space Invasion to use it. Top speed of aliens is 25% higher now, although that could be changed, I kinda liked it. Change parameter order of div().

Some improvements to pas2c

Fix a remote crasher

- Fix some AI stupidness factors - Slight optimizations and formatting

Make frontend compile with clang

- Allow switching colors with mouse wheel - Use indices instead of color values - Introduce a model for colors - Some small fixes, .pro can be used to build again - Yay, tested it to work via network (and fixed a bug!) (TODO: pass colors to engine by index)

initial GLSL1.2 test version

Minor freezer changes, allow snow to accumulate on indestructible terrain too.

Removed lazy update: workaround for R7103

Bring leftTeams back (with a fix) as it is apparently needed for spectators.

merge

- Restore pre-r9257cf8e7af2 behavior - Get rid of leftTeams, send REMOVE_TEAM immediately instead

Merge

Fix reported in bug #327, comment #3 - untested but seems reasonable/safe.

http://www.demoties.com/wp-content/uploads/2011/03/mar-01-6.jpg

Mark team for deletion when letting others know it has left during game. Should help with quitters problem.

Don't remove already removed team

wops add this too

reimplemented ResetKbd and calling it when exiting gsChat status, this restores behavior from.17: You can now walk -> press t/chat -> type some stuff while still walking -> exit gsChat and stop walking at that moment

android: fix the default make target

Android: add freezer commit message

add NOPNG to cmake to explicitly disable PNG dependency reorganise osx detection and variable initalization (simpler crosscompiling) in cmake

add GHC and GHFLAGS from cmake too

rename a few internal cmake variables (for consistency and readability)

Replaced matrix related FFP code with explicit matrix calculations. The modelview matrices now get uploaded at a single point via glLoadMatrix4 rather than using the GL1 matrix stack. In the GL2 variant this will become a uniform upload.

Merge

there a libSDLMain.a in the bundled SDL framework

Merge

android: possible fix for bitmap recycle errors

fixed wheelup/down Wheel up and down used to be swapped in both binds.cpp and uKeys.pas (2 wrongs make a right eh) After I fixed uKeys zooming was inverted, so I changed binds.cpp too, but now existing teams will still have its zoom inverted..

add a nilcheck

checksum

Feel free to add real checksum function

better lazy loading for cScaleFactor

Ok. I think *this* should flip it correctly...

android: added an unsigned target to the makefile, to help automate signing

remove the need to qt and co when using cmake with -DANDROID=1

Minor tweaks to freezer, mostly to simplify current state to laptop

clean up pngs

merge

+ Trivial update. + Script for managing WxW, AFR, and Surf'n'Crate shoppa-variants.

A few scripts to try out. 2x challenge, 1x GSoC training, 1x user mission.

Bit more freezer.

Add an emergency return to the timebox in the case of death of rest of team. Also add a small visual effect to AI survival

Might help if this was actually its own unique weapon

Provide training/campaign with a random seed to ignore/use.

remove ProcessKdb references all together, removed movecursor() because it doesnt seem to do anything

first attempt at fixing controller/gamepad, it works with the android controller but it doesnt have hats/buttons, needs testing

dont allow choosing weapons when the AI is playing

android typo/sabotage

fix sdl13 build

readded optimization for SetScale

unbreak android build - export the Game function - fill the ammo store with 0s in the frondend, if we're missing any weapons - use english as default language

Replaced FFP code that sets up orthogonal projection with explicit matrix calculation. Storing the orthogonal projection in the projection matrix for legacy GL (1.x) rather than in the modelview matrix.

android: add missing graphics, only placeholders tho..

Refactored TTexture to allow encoding sprite(s) at an arbitrary location within a (possibly shared) physical texture. The original variant only allowed encoding a sprite at the bottomleft of a larger physical texture to use NPOT sprites on POT only supporting systems.

Learn hask^Wpascal, dudes!

AI can't use cWindSpeedf since it now does a smooth transition

oops

Dammit Xeli

'\xd7af' -> 0xd7af

Make uUtils.c compile

help pas2c

More magic

unc0rr's magic macro

oops, wrong type

pas2c stuff again

Get rid of cripple SendIPCc

A solution to char vs string problem: mark single-letter strings with _S macro

Make pas2c even more happier with uGears.c, allow assigning arrays in some cases

Add missing identifiers

Fix casing and type for gl

sysutils -> SysUtils

Recognize length on arrays as a separate function

More tolerance to pas2c

Properly track type in ref2CF, this fixes issues with functions returning strings used in expression (like "a" + line())

Define NOCONSOLE for pas2c

uCommandHandlers now compiles

Support operator 'in', replace it with equality checks against each element of set

Recognize (hwFloat op int) expressions

Convert "unit.function()" properly

Function to compare string to char ftw

Make pas2c happier with uSound

Be more wise about constant initialization expressions being not arrays

Avoid returning an array for the sake of pas2c

fix building with lua (i think)

minor changes for warnings and a variables scope

the ios port runs again, although with a few things to sort out

ios: this commit just removes the overlay class files (with all submenus and resources) while working on a proper restoration

Fix identifiers case in uLandTemplates (too lazy to make proper implementation in pas2c)

Declare Param* functions

Fix for Int type and sysutils in hwengine.pas

Some work to make more units compile after conversion to c

Support overloaded operators on (hwFloat op hwFloat) calls

More definitions and mess with pascal code

Render function variables in struct with no mangling. 13 C units are compilable now.

Handle function type differently

Some more definitions and slight fixes

uAmmos.c now compiles (3 warnings though)

Implement built-in functions Low() and High()

Reflects change in previous revision to uTypes

(PointerTo (SimpleType _)) could be a pointer to a non-struct type

Prepend unit name to function identifiers

Mangle overloaded functions (only different number of parameters is supported)

when Lua is not found, fallback to compiling the one that comes bundled in our sources

invert LUA symbol meaning and avoid hints when compiling without it

add support for FPC and FPFLAGS cmake arguments (helps crosscompiling)

old typed const moved to their proper unit

the scope cleanup continues...

split OnDestroy across the appropriate modules (this doen't cause leaks on mobile, right?)

revisit isTerminated in hwengine

fix more breakage

removed the default argument in uWorld

fix breakage

adjust the scope of a few uSound variables and functions

Merge

Convert namespace from list into map in preparation for implementation of overloaded functions support. Greatly improve speed of rendering as a side effect (parse + render time reduced from 5:20 to 0:20)

some tweaks/fixes/etc to the map preview and reload stuff also fixed an old bug of a "stuck random/drawn map preview" when changing maps quickly

visually indicate when map preview generation is in progress.

HAPPY 7000th COMMIT HEDGEWARS - enforce strong typechecking for strings - disable embedded fpc exceptions (binary size savings of ~5k) - add stack check in debug mode

add DLC button to net game room

+ display lack of map in a nice way, don't generate bogus preview in the background * improve map selection behavior in online mode, so that players can load a missing map and have it correctly selected after pressing F5 without any rejoin, map-reselection. also doesn't cause "different maps generated" anymore if map is missing, instead the person lacking the map gets "invalid genmap" and drops out of game immediatly

Get rid of overloaded functions in uRenderUtils

merge

small tweaks/fixes to room list header saving and restore... setModel should be really done in the constructor of the page IMO btw

Switch effects to longint for convenience of tracking ice states. I could add a new Hedgehog value, but since we have this effects list being all useless as booleans anyway...

Updated android support library to the latest version (r8)

fix a previewmap-related QObject::disconnect segfault

Initial stub for freezer

Merge

fix: issue with playing team list vertically collapsing

Workaround for stuck pages.

tweak how the header customization is stored in the settings file

Render float consts into defines

GetRandom -> GetRandomf

allow F5 magic to refresh demos/saves list too

rename a few Draw* routines

wut?

remove old ukeys hooks from pascalexport (definitely breaking the ios port, will have to work on that, but for now sheepluva and xeli will be happy *hopefully*)

rooms list + reorder filter chain for better performance (sorting and full-text search are now last step, not first) + automatically restore custom column widths and sort settings...

roomslist: reimplement filters implementation details: YO DAWG, I herd you like QSortFilterProxyModels, so I put a QSortFilterProxyModel in your QSortFilterProxyModel in your QSortFilterProxyModel in your QSortFilterProxyModel so that you can filter while you filter while you filter while you filter!

remove IFDEF 0 for sake of poor parser

room list: replace magic table column indexes with enum. makes future changes to the room list format way easier.

marked a couple of functions as inline resolved all warnings/hints/notes added -Ooregvar optimisation (stores frequently used variables in registers, saving ~30k code size)

make search filter field have a useable size on low frontend resolution