(svn r7565) -Codechange: Rework DEBUG functionality. Look for appropiate debugging levels to
use in debug.h. grfmsg() is now used as a specific debug-function for grf.

(svn r7313) -Codechange: Calling invalidate data on a window will surely warrant a redraw, so call
that after the WE_INVALIDATE_DATA event and remove (some of) the superflouous calls.

(svn r7195) -Feature: [FS#297, optional elrails] New patches/vehicles option 'disable electrified railways'.
(original patch by maedhros, ideas: peter1138, Darkvater, Rubidium, Patrick, Eddi|zuHause, ..)

(svn r7127) -Codechange [FS#149]: Improve the usability of the signal-dragger, do not bail out
at (certain) errors, just silently ignore them.

(svn r7106) -Feature: Allow over-building of compatible railtypes, i.e. normal and
electrified rail. If building electrified rail, normal rail is upgraded
for you (for the normal cost).

(svn r7048) -Cleanup: DEBUG doesn't need a newline character, use DEBUG instead of printf and the
verb is build/built/built not build/*/build or build/*/builded.

(svn r6774) -Codechange: Rename the badly named OWNER_SPECTATOR to PLAYER_SPECTATOR and
put it into player.h where it belongs (instead of map.h)

(svn r6714) -Codechange: replaced a direct manipulation of windows with InvalidateWindowData() in rail_cmd.c
Moved the actual modification of railtype to WE_INVALIDATE_DATA in the train depot handler
-Codechange: added SetWindowDirty() to WE_INVALIDATE_DATA as it made no sense to update the list without making the window dirty

(svn r6651) -Coding feature: added the windowevent WE_INVALIDATE_DATA
This gives the ability to invalidate some window data and recalculate as needed instead of doing it for each WE_PAINT

This event is called right away when using InvalidateWindowData(), so it may be a good idea to set a bool or similar in the window
or similar and then act on that bool in WE_PAINT instead of doing a lot of stuff in WE_INVALIDATE_DATA as it might be called more than once before WE_PAINT is called

InvalidateWindowData() will not automatically repaint the window, so if you want to repaint it as well, you need to mark it dirty as well.

Made the depot windows use WE_INVALIDATE_DATA to set when to generate the engine and wagon lists instead of at each redraw
It makes no sense to regenerate the list when say using the scrollbar if we know that no vehicle have entered or left the list

NOTE: currently there is a piece of code to generate the list when it's not needed and compare it to the stored list and assert if they mismatch
This check is somewhat slow and kills the whole idea of WE_INVALIDATE_DATA, so it's a short lived one to verify that InvalidateWindowData() is used everywhere where it's needed

(svn r6637) -Codechange: merged all (vehicle type)EnterDepot into VehicleEnterDepot()
This revealed duplicated code like aircraft lists got invalidated twice
Moved invalidation of the vehicle detail window to VehicleServiceInDepot() as it should always be updated when serviced

(svn r6513) -Codechange: unified the code to draw depot windows
This change is intended to make it easier to make depot behaviour consistent
and faster to code when adding more features in the future
The user interface should hopefully not be affected by this

(svn r6406) -Codechange: Rename TileOffsByDir to TileOffsByDiagDir because it accepts
DiagDirections, and add TileOffsByDir that handles Directions.
-Codechange: Make the treeloop use TileOffsByDir().

(svn r6381) -Cleanup: make the '/* */' comments that span multiple lines more uniform.
-Cleanup: whitespace alignment of a few tables.

(svn r6141) -Codechange: introduced DepotID and used it as much as possible
-Codechange: DeleteDepot removes a depot from the pool
-Codechange: DestroyDepot is called by DeleteDepot to remove all things where a depot depends on.
Last 2 changes to prepare for new pool system. Not pretty now, will be soon.
-Codechange: Removed DoDeleteDepot as it was stupid

(svn r5396) - Remove two fixed parameters
- Simplify an expresssion: c ? f(x) : f(y) -> f(c ? x : y)

(svn r5363) Revert 5312, 5288, 5248, 3992, 3249, 3228
They were all aimed at fixing the terraform-into-tunnel problem, but introduced new sideeffects while doing so

(svn r5326) Remove the indicator bit for custom waypoint graphics, because it just contains redundant information
Yup-yup-ed by: peter1138

(svn r5315) -Fix: Prohibit altering a road tile while road works are in progress
This fixes some glitches like "turning" the excavation by adding/removing road bits or removing the road piece

(svn r5070) Merged the bridge branch
-Feature: Bridges can now be placed above:
Any railway track combination (excluding depots and waypoints)
Any road combination (excluding depots)
Clear tiles (duh), including fields
Tunnel entrances
Bridge heads

Thanks to Tron for idea and implementation, KUDr for the yapf synchronization and many others for hours of testing

There are still a number of visual problems remaining, especially when electric railways are on or under the bridge.
DO NOT REPORT THOSE BUGS FOR THE TIME BEING please.

(svn r4593) CodeChange : Renamed spec (when using StationSpec) to statspec.
This is following the same scheme as for IndustrySpec

(svn r4572) Remove vehicle_leave_tile_proc
There's exactly one implemented function (unbarring a level crossing) and one place where it can be triggered (a train leaves a level crossing)
It's a bit overkill, so just handle this one case where it can happen

(svn r4546) - NewGRF: add support for Action 0x05, type 0x04: replacement signal graphics. Thanks to Purno for supplying data to test
with.

(svn r4415) Fix bug introduced in r4384 where drawing a default waypoint for the GUI
caused a game crash. This only manifested with some newstations loaded.
Thanks to Richk67 for finding this.

(svn r4384) -Codechange: rail drawing code
-Added a function to obtain the axis of a waypoint
-Moved the waypoint layout tables from the depot layout tables into an own one
-Use GetRailTileSubtype

(svn r4342) Change the first two parameters of commands - virtual pixel coordinates of the tile to operate on - to a TileIndex
Remove DoCommandByTile(), because now it does the same as DoCommand()