summaryrefslogtreecommitdiff
path: root/src/yapf
AgeCommit message (Collapse)Author
2007-07-29(svn r10733) -Codechange: change MP_STREET into MP_ROAD as we use the word ↵rubidium
"road" everywhere except in the tile type.
2007-07-15(svn r10578) -Fix [YAPF, ships]: Ships received curve penalty for ↵KUDr
non-diagonal straight move. (JazzyJaffa) -The fix in cost calculation uncovered bug in estimate calculation. Ships now use the same estimate algorithm as trains.
2007-07-09(svn r10492) -Fix [YAPF, r10491]: commands with no effect (glx)KUDr
2007-07-09(svn r10491) -Fix [FS#988, YAPF]: When rail segment was cached using ↵KUDr
electric engine and it ended with non-electric rail it was incorrectly cached with the end reason ESRB_DEAD_END instead of ESRB_RAIL_TYPE. (Eddi) - It caused YAPF to end prematurely there when it was searching for another path (for non-electric engine). - It can lead to sub-optimal path taken or 'train is lost' message. - In MP game it can also cause desync. - Uses changes made in r10489.
2007-07-09(svn r10489) -Codechange [YAPF]: FollowTrack_t now has new data member - ↵KUDr
m_err. It indicates the reason why the given track/trackdir can't be followed when method CFollowTrackT<>::Follow() returned false.
2007-07-04(svn r10423) -Fix [YAPF, r10301]: INVALID_TRACKDIR causing buffer overrun ↵KUDr
leading to desync in MP game (many thanks Rubidium)
2007-06-30(svn r10395) -Fix [YAPF]: missing include + some signed/unsigned comparison ↵KUDr
warnings
2007-06-29(svn r10394) -Fix [YAPF]: MP desync caused by incorrect last-red-signal ↵KUDr
penalties applied when cached segment was reused. Players that connected more recently (didn't have these segments cached yet) calculated those penalties correctly. This caused different YAPF results.
2007-06-29(svn r10393) -Fix [YAPF]: Trackdir needs 4 bits to store (not 3). The cached ↵KUDr
segment key didn't use the highest trackdir bit. It could confuse YAPF on bi-directional track segments.
2007-06-29(svn r10392) -Add [YAPF]: added structured dump support into some essential ↵KUDr
YAPF classes (node-list, nodes, keys, etc.) and CArrayT
2007-06-27(svn r10366) -Codechange [YAPF]: added DEBUG_YAPF_CACHE macro that (when set ↵KUDr
to 1) allows to track YAPF "cache errors". They are probably responsible for current MP desyncs. (thanks Rubidium for this great idea!). I will deal with those errors/desyncs tomorrow.
2007-06-26(svn r10349) -Fix [FS#941, YAPF, r10301]: tile/trackdir must be used as node ↵KUDr
key and also as segment key in the cache (SmatZ). In the past it was possible to use tile/exitdir as the key because segments beginning on the same tile/exitdir were incorrectly considered the same. What I still don't understand is why this bug happened only on 64 bit systems (linux, Win64).
2007-06-24(svn r10304) -Fix [YAPF](r10301): removed assert which could violate without ↵KUDr
danger when searching for the nearest depot. From now on the segment can be closed without reason in that case (peter1138)
2007-06-24(svn r10303) -Fix [YAPF](r10301): warning: unused variable tile_cost (peter1138)KUDr
2007-06-24(svn r10301) -Fix [FS#901, YAPF]: another assert violation in some special ↵KUDr
cases (immeR)
2007-06-18(svn r10199) -Fix [FS#884, YAPF]: 'target_seen' flag that is set prematurely ↵KUDr
in some cases (1 tile long cached segment followed by target station) stops on assert (immeR)
2007-05-31(svn r10006) -Fix (FS#362): allow different signal types on one tileglx
2007-05-25(svn r9922) -Fix (r9914): [YAPF] null pointer dereference.rubidium
2007-05-24(svn r9914) -Codechange: prepare GTTS and the pathfinders to handle multiple ↵rubidium
road types on a single tile.
2007-05-18(svn r9869) -Codechange: replace some bytes with VehicleType, i.e. more type ↵rubidium
strictness.
2007-04-29(svn r9746) -Codechange [YAPF]: added DEBUGs level 0 if the current train ↵KUDr
order is OT_GOTO_WAYPOINT and waypoint is invalid or v->dest_tile is wrong
2007-04-29(svn r9745) -Codechange [YAPF}: if destination is a waypoint then YAPF gets ↵KUDr
destination tile from that waypoint->xy instead of from v->dest_tile
2007-04-20(svn r9697) -Fix [YAPF](r9694): 'unused variable' warning (glx)KUDr
2007-04-20(svn r9694) -Fix [YAPF][FS#736]: the guessed path (when PF stopped on max. # ↵KUDr
of nodes passed) was ignored for ships (desolator)
2007-04-20(svn r9693) -Codechange [YAPF]: GetBestNode() now returns pointer to node ↵KUDr
instead of reference
2007-04-18(svn r9665) -Documentation: Doxygen corrections,errors, corrections of ↵belugas
corrections...
2007-04-12(svn r9609) -Codechange: Move some function prototypes out of functions.h ↵maedhros
and into landscape.h, and add a few where they didn't exist.
2007-03-08(svn r9068) -Codechange: capitalize the VEH_Train etc. enums to match the ↵rubidium
coding style (and rest of the code).
2007-03-07(svn r9051) -Codechange: typedef [enum|struct] Y {} X; -> [enum|struct] X {};rubidium
2007-03-07(svn r9050) -Codechange: Foo(void) -> Foo()rubidium
2007-03-03(svn r8987) -Cleanup: doxygen changes. Again. Mostly (still) @files missing ↵belugas
tags and (more than just) a few comments style.
2007-02-25(svn r8899) -Fixtron
Remove the unused second parameter from GetTileRailType()
2007-02-24(svn r8869) [YAPF] -Fix: Large Train Stations/Trains makes OpenTTD crash ↵KUDr
(Jigsaw_Psyche)
2007-02-23(svn r8867) -Fix(r8866): revert yapf_costrail.hpp commited by mistake, ↵KUDr
helpers.hpp had wrong line ending
2007-02-23(svn r8866) -Doc: added more comments to template struct MakeEnumPropsTKUDr
2007-02-23(svn r8857) -Documentation: Added some doxygen @file tags, repaired others ↵celestar
(the @file tag MUST be found before any line of code, that includes preprocessor directives).
2007-02-21(svn r8836) [YAPF] -Fix[FS#641]: Assertion: 'IsTileDepotType(depot_tile, ↵KUDr
TRANSPORT_ROAD)' failed (Karsten) -don't tell that we found destination if the best direction was only guessed
2007-02-17(svn r8766) -Fix[YAPF]: rail station platform penalty not calculated properlyKUDr
2007-02-14(svn r8735) -Feature: drive-through road stops made possible by the hard ↵rubidium
work of mart3p.
2007-02-13(svn r8709) -Fix/Codechange: Rename the function GetStationPlatforms into ↵celestar
GetPlatformLength because that is what it really does. Overload it because there is already a GetPlatformLength (one gives the length of the whole platform, the other gives the remaining length in a given direction). Turned both functions into methods of Station. While messing around with it, fix a problem where loading times for overhanging trains are miscomputed.
2007-01-24(svn r8392) -Fixtron
Use HASBIT() instead of implementing it manually
2007-01-13(svn r8092) -Codechange: header files with miscellaneous template classes ↵KUDr
(smart pointers, blob, array, hashtable, etc.) moved from src/yapf to src/misc as they can now be used anywhere.
2007-01-12(svn r8080) -Codechange (r8079): Move the *WHOLE* performance code into the ↵Darkvater
#ifndef and some style changes.
2007-01-12(svn r8079) -Fix [YAPF]: float division by zero when calculating stats (YAPF ↵KUDr
cache hit ratio). Caused BSOD on Win9x. (thanks 3iff for report, Darkvater for help)
2007-01-10(svn r8038) -Merge: the cpp branch. Effort of KUDr, Celestar, glx, Smoovius, ↵rubidium
stillunknown and pv2b.
2007-01-02(svn r7759) -Merge: makefile rewrite. This merge features:rubidium
- A proper ./configure, so everything needs to be configured only once, not for every make. - Usage of makedepend when available. This greatly reduces the time needed for generating the dependencies. - A generator for all project files. There is a single file with sources, which is used to generate Makefiles and the project files for MSVC. - Proper support for OSX universal binaries. - Object files for non-MSVC compiles are also placed in separate directories, making is faster to switch between debug and release compiles and it does not touch the directory with the source files. - Functionality to make a bundle of all needed files for for example a nightly or distribution of a binary with all needed GRFs and language files. Note: as this merge moves almost all files, it is recommended to make a backup of your working copy before updating your working copy.