Release date: 05.08.2008
- All
- debugged bug 1976405 Saving invalid dump configuration; Was a false encoding in Option_IntVector
- builds under 64bit Linux (solved bugs bug 1927822, bug 1928370, bug 1930444, bug 1931468, bug 1932961, bug 1933569, bug 1934635; thanks to Ronald Nippold for being persistent in supplying the build logs
- encapsulated xerces-attributes; makes value retrieval so much more comfortable; retrieve a attribute using attrs.getTYPENAME(id) instead of getTYPENAME(attrs, id)
- Simulation
- Hidden state loading/saving completely from the microsim
- Vehicle may now end their route at a certain position of the route's last edge (see Specification)
- The route may be now given as an attribute (route) within the vehicle definition (see Specification)
- The vehicle now stores most of its definition parameters in myParameter, not as atomar values
- Solved problems with multiple stop on same edge and with stops on lanes which can not be used for continuing the route; thanks to Friedemann Wesner and Gerrit Lammert for reporting them
- Reworked mean data output. The vehicle movements are now weighted by the time the vehicle was on an edge/lane. Additional constraints for output (--exclude-empty-edges/--exclude-empty-lanes) allow to dump only information about edges/lanes which were occupied by vehicles.
- debugged bug 1964433 different simulation results for different starting times; The issue was a false computation of WAUT switching and TLS offset times when choosing a different simulation begin time than 0
- replaced MSVehicle::getInTransit and MSVehicle::running by MSVehicle::isOnRoad
- removed MSEdge::isSource; was used only once and can be easily replaced by comparing the result of MSEdge::getPurpose()
- debugged bug 1932105 second stop on the same edge confuses vehicles
- debugged bug 1941846 repeatable crash with sumo r5378
- debugged bug 1912547 linux 64 bit compile error for sumo svn 5143
- debugged bug 1913655 command line sumo throws unknown exception
- debugged bug 1924989 sumo r5186 crashs by pure virtual function call
- begun work on Specification
- all vehicles are emitted at position=0, with speed=0
- removed usage of NLLoadFilter
- added an option to quit the simulation if a number of vehicles is exceeded (--too-many-vehicles <INT>)
- fully integrated Tino Morenz's MSVTypeProbe; moved it from
microsim/trigger to microsim/output; tests added; a vtypeprobe
is now defined within additional files this way:
<vtypeprobe id="test1" type="" freq="10" file="vtypeprobe.xml"/>("test1" is the vehicle type to report, leave empty in order to get information about all vehicles) - replaced usage of "Command" by encapsulating methods in "WrappingCommand" wherever possible
- "debugged" usage of actuated/agentbased tls; should be covered by tests, in fact
- further work on devices-consolidation
- implemented a one-shot / multi-shot routing device
- sumo-gui
- Removed opening an additional command line window from the release version
- Title is now set properly
- debugged bug 2021454 Discarding View settings resets them; The editor now uses the settings that were used when it was opened as the one to apply when "cancel" is pressed
- debugged bug 2021464 view settings are not saved; The button is now removed, all the interaction will be done via the top buttons
- debugged bug 2015792 Save List of selected Items; The proper extension (mostly ".txt") is now appended if no extension is given. Affects saving breakpoints, additional weights, tracked parameter, and selections.
- bug 1919509 guisim crashed is invalid, the issue was a buggy opengl driver
- netconvert
- debugged bug 1993928 --keep-edges.input-file does not remove empty nodes; as soon as one edge-removal(keeping) option occures, nodes are rechecked whether they contain any further edge. If not, the node is removed.
- VISUM-import: debugged failures due to false district node position computation
- Connection building: made storing outgoing edge connections sane; replaced three strange storages by one
- Disabled the support for setting an "edge function" - because vehicles are emitted the same (parametrised) way on each edge, we do not need a distinction between source and normal edges; sink edges were never distinguished.
- Refactored the way edge types are stored in netconvert
- removed "name" attribute from NBEdge; was not written anyway
- work on SUMO-networks reimport (unfinished)
- documentation: debugged bug 1909685 option "omit-corrupt-edges" not recognized (use --dismiss-loading-errors instead of --omit-corrupt-edges)
- removed usage of NLLoadFilter
- thanks to Christoph Sommer, the OSM importer now handles speed and lanes information; default OS values are now used, too
- duarouter / jtrrouter
- Removed option --move-on-short. It got obsolete as vehicles may start on edges which are shorter than they are, now.
- duarouter
- debugged bug 1981921 duarouter truncates routes; Now, --remove-loops only prunes the route if the vehicle would have to take a turnaround at the begin or end.
- reworked handling of broken routes; longer routes are not replaced; incomplete routes may be corrected using the new --repair option
- jtrrouter
- Removed possibility to specify turning percentages using CSV files
- dfrouter
- Definition of emitters, vss, etc. are now written into the same folder as the declaration
- TraCI
- TraCI is now enabled as default
- od2trips
- due to being used for macroscopic demand definitions, od2trips now emits vehicles using departlane="best" and departspeed="max"
- removed option to parse list of tables from a vissim file; they may be given at the command line (--vissim is no longer supported)
- added the possibility to omit writing the vehicle type (--no-vtype)
- Tools
- added some tools for output visualization (mpl_dump_twoAgainst.py, mpl_tripinfos_twoAgainst.py, mpl_dump_timeline.py) (see Tools/Visualization)