Dennis Luxen
|
3db44b3ce1
|
minor changes
|
2011-03-15 15:20:14 +00:00 |
|
Dennis Luxen
|
e695e269b1
|
introducing tuning parameters
|
2011-03-15 15:19:20 +00:00 |
|
Dennis Luxen
|
be34eebda7
|
Writing level information into seperate file
|
2011-03-14 18:01:02 +00:00 |
|
Dennis Luxen
|
d2c532e4d5
|
Removed Dead code
|
2011-03-10 16:01:44 +00:00 |
|
Dennis Luxen
|
e32ba24b6a
|
fixing ticket 20. stxxl should not complain any more. also, there should be no compiler warnings.
|
2011-01-14 16:54:42 +00:00 |
|
Dennis Luxen
|
cf46fd79b0
|
PBF Support (Thanks to Christian for explanations)
|
2011-01-12 18:08:10 +00:00 |
|
Dennis Luxen
|
bfd2a8aee2
|
This is a large update that brings many internal and architectural changes. The most obvious change to the user is the presence of configuration files for extractLargeNetwork and routed. Optimistically speaking, it should not break anything. Thanks to rskr for support patches and suggestions.
|
2011-01-09 21:42:27 +00:00 |
|
Dennis Luxen
|
17310a9d2b
|
Fixes a race condition when more than one thread tries to access the grid file (thanks Patrick)
|
2010-11-18 16:58:31 +00:00 |
|
Dennis Luxen
|
03772fd85b
|
Preprocessing of routing datastructures at least 30 percent faster (Monav backport, thanks Christian)
|
2010-11-17 14:03:18 +00:00 |
|
Dennis Luxen
|
0f1f99dbf2
|
Propagating turn info to hsgr
|
2010-10-04 11:00:36 +00:00 |
|
Dennis Luxen
|
5692d43abb
|
Propagating turn info (not yet finished)
|
2010-10-01 16:30:35 +00:00 |
|
Dennis Luxen
|
3b885a492c
|
Every node in route carries a bit that indicates if its possible at all to take a turn
|
2010-10-01 10:32:39 +00:00 |
|
Dennis Luxen
|
e91058c2eb
|
Output formatting
|
2010-10-01 10:30:46 +00:00 |
|
Dennis Luxen
|
5637c2f92a
|
BREAKING CHANGE, REPROCESS YOUR OSM FILES
Information if its possible at all to turn at the end of edges is propagated through the complete toolchain
|
2010-09-30 15:40:48 +00:00 |
|
Dennis Luxen
|
94cfb4aad7
|
nompute a vector<bool> that indicates if it's possible at all to take a turn at a certain node
|
2010-09-29 16:40:36 +00:00 |
|
Dennis Luxen
|
676f64b0ef
|
BREAKING CHANGE, REPROCESS YOUR OSM FILES
All preparations necessary to compute and output turn directions.
|
2010-09-29 15:22:38 +00:00 |
|
Dennis Luxen
|
4c47d5b70e
|
segfault if first routing node equals start node
|
2010-09-29 13:22:12 +00:00 |
|
Dennis Luxen
|
76488c58e8
|
Preparing data structures for turn directions
|
2010-09-23 15:34:22 +00:00 |
|
Dennis Luxen
|
586e8b4c4b
|
Saving a byte per node
|
2010-09-22 10:21:18 +00:00 |
|
Dennis Luxen
|
2acfe2591f
|
Fixing a rare case of route construction problems when origin were on the same edge which was stored reversed in graph.
|
2010-09-22 07:50:00 +00:00 |
|
Dennis Luxen
|
8f06b2fa31
|
It is way too late to fix silly bugs.
|
2010-09-21 17:15:59 +00:00 |
|
Dennis Luxen
|
c36fc9e49f
|
Forgot to guard against bad data from NNGrid
|
2010-09-21 16:23:50 +00:00 |
|
Dennis Luxen
|
ee56d0051f
|
Fixing odd behavior when origin and destination are on the same street.
|
2010-09-21 15:59:52 +00:00 |
|
Dennis Luxen
|
61a8b6820f
|
Same include twice
|
2010-09-20 09:32:16 +00:00 |
|
Dennis Luxen
|
be836dc950
|
fixing last leg of output path
|
2010-09-17 08:46:43 +00:00 |
|
Dennis Luxen
|
41e5a52509
|
last leg of route is now computed correctly
|
2010-09-16 15:08:45 +00:00 |
|
Dennis Luxen
|
a07efcc4b7
|
moving common code into a single file
|
2010-09-15 13:49:26 +00:00 |
|
Dennis Luxen
|
94c8c61a05
|
New auxiliary method
|
2010-09-14 16:40:11 +00:00 |
|
Dennis Luxen
|
43bbf953c1
|
Fixes bad alloc for large graphs
|
2010-09-13 15:31:29 +00:00 |
|
Dennis Luxen
|
1ba915cbed
|
BREAKING CHANGE, REPROCESS YOUR OSM FILES
Ferry egdes are now ignores by nearest neighbor grid
|
2010-09-02 15:47:55 +00:00 |
|
Dennis Luxen
|
e79e36bea9
|
Perfomance fixes
|
2010-09-02 13:01:49 +00:00 |
|
Dennis Luxen
|
0875c2d557
|
BREAKING CHANGE: the routing engine does not use the kd-tree anymore, but an external memory nearest neighbor grid. The memory consumption is less than half. This is a breaking change, please reprocess your .osrm files and take note that the parameter to routed have been changed!
Known Bug: There are some minor rounding errors that affect the position of start and destination node by a few feet.
|
2010-08-31 14:00:40 +00:00 |
|
Dennis Luxen
|
b3fded0725
|
Reverting change that made path unpacking impossible
|
2010-08-25 09:33:41 +00:00 |
|
Dennis Luxen
|
a9b76e8050
|
Using less RAM, Fixing a leak in ContractionCleanup.
|
2010-08-24 13:50:58 +00:00 |
|
Dennis Luxen
|
106155a97a
|
forgot debug output
|
2010-08-12 15:43:49 +00:00 |
|
Dennis Luxen
|
59abeccc3e
|
Show the number of connected components and the size of the largest one
|
2010-08-12 15:42:22 +00:00 |
|
Dennis Luxen
|
d07dd71078
|
Dense is dense and sparse is sparse
|
2010-08-07 12:25:45 +00:00 |
|
Dennis Luxen
|
051b710484
|
Use different map
|
2010-08-06 16:24:06 +00:00 |
|
Dennis Luxen
|
ec0108553f
|
removed libkdtree++ dependency.
|
2010-07-26 08:17:52 +00:00 |
|
Dennis Luxen
|
a83b790225
|
GCC 4.2 Compile fix. Code compiles with older GCC, that the default compiler on Mac OS. Only a few warnings left.
|
2010-07-24 14:05:03 +00:00 |
|
Dennis Luxen
|
72e314d1c0
|
kdtree usage simplified and dropped intermediate data structures
kdtree build-up twice at fast
int2ext node map holds iterators instead of object copies
some functions const'ed
buggy defines fixed
|
2010-07-14 16:29:18 +00:00 |
|
Dennis Luxen
|
fa5a042071
|
Queries are now handled by a static graph which means less RAM usage and faster queries.
|
2010-07-14 12:55:53 +00:00 |
|
Dennis Luxen
|
d4a64d2168
|
Initial Import.
|
2010-07-09 09:05:40 +00:00 |
|