summaryrefslogtreecommitdiff
path: root/src/pathfinder
AgeCommit message (Collapse)Author
2018-10-31Codechange: Some more null checksCharles Pigott
2018-10-31Fix: Protect against a few out of bounds or uninitialised usage errorsCharles Pigott
2018-10-31Doc: Lots and lots of doxymentation fixesCharles Pigott
2018-06-27Codechange: Use HasTrack(dir) to improve code readability.J0anJosep
2018-06-27Codechange: Increase readability of track functions and pathfinders.J0anJosep
2018-06-27Cleanup: Unnecessary assignation on FollowTileExit(): done previously on the ↵Juanjo
function.
2018-06-27Codechange: Separate an assertion. More information if assert is triggered.Juanjo
2018-04-30Codechange: Use TileAddBy(Diag)Dir when possible.J0anJosep
2018-04-11Remove: NO_DEBUG_MESSAGES was only read and setting it broke compilation (#6703)Patric Stout
Given any speed issue cannot be attributed to checking for _debug_NNN_level, removing this is a safe action This fixes #6652.
2017-09-03(svn r27912) -Fix (r13948): [NPF] Reserved track bits were not accounted for ↵michi_cc
when trying to find any safe position.
2017-08-31(svn r27906) -Cleanup: Remove some NPF code with no effect and mark possible ↵adf88
bug about never used NPF_FLAG_IGNORE_RESERVED flag
2017-08-20(svn r27896) -Fix: some warningsfrosch
2017-08-13(svn r27893) -Codechange: Use fallthrough attribute. (LordAro)frosch
2017-06-22(svn r27885) -Cleanup (r27844): ESR_MAX_COST_EXCEEDED and ESR_PATH_TOO_LONG ↵frosch
meant the same thing, but only one was used. Keep PATH_TOO_LONG since it has the better documentation.
2017-04-04(svn r27846) -Fix [FS#5926]: Infinite loop in pathfinder when checking safe ↵peter1138
waiting position from a waypoint.
2017-04-03(svn r27845) -Change: (Yapf) Use FindDepotData struct to simplify depot ↵peter1138
finding code and remove need to return fake path distance. (juanjo)
2017-04-03(svn r27844) -Change: (Yapf) Treat max cost exceeded separately from path ↵peter1138
too long condition, as destination should not be considered in the former case.
2017-04-03(svn r27843) -Change: (Yapf) Consider depot as destination before reversing ↵peter1138
path and applying penalty.
2017-01-15(svn r27733) -Codechange: Pass NPF user data as void* instead of as array of ↵frosch
integers.
2016-05-29(svn r27586) -Fix [FS#6410]: Automatic servicing of road vehicles compared ↵frosch
path finder costs with tile distances, thus vehicles went to depots which were factor 100 too far away. (juanjo)
2015-10-30(svn r27418) -Fix [FS#6329] [FS#6379]: Desync due to incorrect storage of ↵frosch
segments with different railtype in the YAPF cache. (JGR)
2015-08-08(svn r27363) -Codechange: Fix codestyle of one-line methods and header ↵alberth
codestyle of derived structs.
2015-08-08(svn r27362) -Codechange: Codestyle fixes for reference var declarations, ↵alberth
static cast type, operator methods.
2015-07-12(svn r27332) -Change: Enable YAPF cache debugging with desync debug level 2.frosch
2015-01-02(svn r27107) -Fix [FS#6183]: road vehicles could not reverse to be sent to ↵rubidium
depots when the following tile has the right type to run on, but could not be entered; e.g. facing toward a depot with the entry facing the wrong way failed to reverse whereas facing towards an empty tile allowed the vehicle to reverse (estys)
2014-04-23(svn r26482) -Codechange: add an include that allows us to undefine/redefine ↵rubidium
"unsafe" functions to prevent them from being used, and thus having to care about certain aspects of their return values
2013-11-26(svn r26117) -Cleanup: unneeded NULL checkrubidium
2013-11-25(svn r26109) -Fix: missing initialisationrubidium
2013-11-25(svn r26102) -Fix: m_veh is not NULL by definition when IsRoadTT is truerubidium
2013-11-23(svn r26058) -Fix: handle the return value of a number of functions betterrubidium
2013-07-14(svn r25609) -Fix [FS#5216]: under certain circumstances a track type change ↵rubidium
would make the end-of-line-is-red setting ineffective
2013-07-14(svn r25608) -Codechange: make EndSegmentReasonBits available in yapf_base.hpprubidium
2013-02-03(svn r24964) -Fix [FS#5416]: With YAPF the docking behaviour differed per ↵rubidium
direction; now favour docking in the direction you approached
2013-01-11(svn r24905) -Feature(ish): Implement station randomisation triggers.peter1138
2013-01-08(svn r24900) -Fix [FS#5389]: Comments with typos (most fixes supplied by ↵planetmaker
Eagle_rainbow)
2012-10-13(svn r24588) -Codechange: No need to intermingle reservation-checks with ↵frosch
railroad checks. (Juanjo)
2012-10-13(svn r24587) -Codechange: Improve coding style (Juanjo)frosch
2012-08-18(svn r24481) -Feature [FS#5127]: Make the pathfinder decide whether ships ↵frosch
shall leave depots towards north or south.
2012-08-18(svn r24479) -Fix: Trains were unable to reverse in stations when using NPF.frosch
2012-01-03(svn r23740) -Codechange: remove some 300 unneeded includes from the .cpp filesrubidium
2011-12-20(svn r23640) -Fix: stop using FORCEINLINE (1/3rd of the instances were, the ↵truebrain
others were still regular inline), but make sure inline is always a 'forced' inline (I am looking at you MSVC)
2011-12-17(svn r23564) -Fix [FS#4888]: Extending a path reservation starting at a ↵michi_cc
partially reserved rail station could fail.
2011-11-30(svn r23376) -Remove: on popular demand, remove my (revoked) name from ↵truebrain
comments in the code. It was silly to name me like that to begin with ;) (based on patch by HackaLittleBit)
2011-11-04(svn r23093) -Codechange: add a default NULL for the Z of GetTileSlope and ↵rubidium
use it
2011-11-04(svn r23091) -Codechange: rename some Get*Z functions to Get*PixelZ ↵rubidium
functions if they return the Z in pixels (like TilePixelHeight)
2011-09-03(svn r22881) -Fix (r22875): GCC warnings on 64bit systems.frosch
2011-09-02(svn r22875) -Codechange: Add some asserts and checks to better prevent ↵michi_cc
overflow of the argument to malloc. (monoid)
2011-08-21(svn r22801) -Feature: [YAPF] Take canal/ocean speed fraction of ships into ↵michi_cc
account.
2011-05-02(svn r22410) -Document: some more bits ;)rubidium
2011-04-19(svn r22350) -Change: make YAPF hash tables much smaller, it should improve ↵smatz
performance significantly (about 15%)