[mkgmap-dev] spltter software design question
From GerdP gpetermann_muenchen at hotmail.com on Sun Oct 21 11:30:16 BST 2012
Hi, I am still working on the cleanup of the patch for splitter (splitter_problem_list.patch) I got the impression that the patched code is quite messed up now regarding the naming and use of interfaces and abstract classes. I'd like to change that so that a) - MapReader defines the interface to the IO routines that read OSM data, each format like pbf, o5m, or osm(xml) requires a special MapReader implementation. Current r202 defines an empty and more or less unused interface - MapProcessor defines the interface to the routines that should process the OSM data provided by the IO routines (nodes, ways, relations, etc). - OSMWriter defines the interface to the IO routines that write OSM data, each format like pbf, o5m, or osm(xml) requires a special OSMWriter implementation. Current r202 defines this as an abstract class. - class AbstractMapReader should contain common code in (at least two) readers - class AbstractMapProcessor should contain common code in (at least two) processors - class AbstractOSMWriter should contain common code in (at least two) writers b) The NodeCollector (used with parm --legacy-mode) seems to be obsolete to me, I want to remove it c) A few other helper classes are no longer in use (e.g. IntList, IntObjMap, SparseInt2ShortMap). I'd like to move them to a folder called obsolete What do you think about this? -- View this message in context: http://gis.19327.n5.nabble.com/spltter-software-design-question-tp5732135.html Sent from the Mkgmap Development mailing list archive at Nabble.com.
- Previous message: [mkgmap-dev] What the "crosby.binary.file.FileFormatException: ... Possibly corrupt file." means in splitter
- Next message: [mkgmap-dev] spltter software design question
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the mkgmap-dev mailing list