December 2008 - mkgmap-dev - The mkgmap lists

more public testing?
by Robert Vollmert 06 Dec '08

06 Dec '08

Hi, routing now seems to work well enough on my Vista. What bugs are left have no noticeable or clear effect, so I'm having trouble testing. But I also have no experience how routing should work with a proper map. What do you think of trying to get some more people to try it out in hope of getting more precise information on what doesn't work yet? I.e., post to talk/wiki/forum asking people to try out the nod branch. I'm hoping other Garmin products will use the map differently and fail in helpful ways. Here's what's still missing as far as I see it: * Not writing any links. * Not writing curve data. * Possibly not writing the correct destination class on arcs. * Not writing road length to net (there's uncertainty as to the unit). * Writing arc length to NOD 1, put possibly the same uncertainty as to unit? * Not writing boundary node data (NOD 3). Does anyone know how polish format handles boundary nodes? Is it one .mp-file per IMG and boundary nodes are marked with flags? Or is it one .mp-file and the compiler handles the cutting up? Cheers Robert

2 4

Commit: r777: A little more relaxed with respect to the input.
by svn commit 06 Dec '08

06 Dec '08

Version 777 was commited by robert on 2008-12-06 12:04:00 +0000 (Sat, 06 Dec 2008) A little more relaxed with respect to the input. * If length is too large for our encoding, truncate and warn instead of asserting. * Don't try to write to Nod 2 for roads without routing nodes.

1 0

Commit: r776: Make the MPS file name distinctive.
by svn commit 06 Dec '08

06 Dec '08

Version 776 was commited by steve on 2008-12-06 11:42:37 +0000 (Sat, 06 Dec 2008) Make the MPS file name distinctive.

1 0

Commit: r775: Calculate the block size needed for the gmapsupp correctly. Was ignoring the main constraint
by svn commit 06 Dec '08

06 Dec '08

Version 775 was commited by steve on 2008-12-06 09:42:55 +0000 (Sat, 06 Dec 2008) Calculate the block size needed for the gmapsupp correctly. Was ignoring the main constraint while concentrating on a minor one.... Also there should be a warning now if the directory overflows. If it does that would be a bug.

1 0

Commit: r774: Read boundary bit in .mp and write NOD3.
by svn commit 06 Dec '08

06 Dec '08

Version 774 was commited by robert on 2008-12-06 09:32:12 +0000 (Sat, 06 Dec 2008) Read boundary bit in .mp and write NOD3.

1 0

Commit: r773: Prepare for writing curve data. Cosmetics.
by svn commit 06 Dec '08

06 Dec '08

Version 773 was commited by robert on 2008-12-06 08:50:45 +0000 (Sat, 06 Dec 2008) Prepare for writing curve data. Cosmetics.

1 0

Commit: r772: Rewrite arc length handling. Also calculate road length.
by svn commit 05 Dec '08

05 Dec '08

Version 772 was commited by robert on 2008-12-05 17:13:37 +0000 (Fri, 05 Dec 2008) Rewrite arc length handling. Also calculate road length. Arc length now uses one specific encoding with curve=false and 14 bits for the length. RouteArc.write is prepared for other encodings. (NodDisplay also knows encodings with 10 bits and 22 bits (last only if curve=true).) RoadDef now gets passed the road's entire length for writing to NET, though we still write a hard-coded value.

1 0

Commit: r771: Now computing real arc length in RoadNetwork, but not writing it yet.
by svn commit 05 Dec '08

05 Dec '08

Version 771 was commited by robert on 2008-12-05 15:17:18 +0000 (Fri, 05 Dec 2008) Now computing real arc length in RoadNetwork, but not writing it yet. Some cosmetic fixes. Moved distance calculation from RouteArc to Coord -- or should it go in Utils?

1 0

Commit: r770: Complete code for writing NOD3.
by svn commit 05 Dec '08

05 Dec '08

Version 770 was commited by robert on 2008-12-05 12:05:51 +0000 (Fri, 05 Dec 2008) Complete code for writing NOD3. imgfmt.app.net is now NOD3-ready; RoadNetwork will need to be updated to create boundary nodes for nodes on the boundary (however that will be defined, probably through the clipper).

1 0

Commit: r769: Only write extra bits if there are internal nodes (not sure this is
by svn commit 05 Dec '08

05 Dec '08

Version 769 was commited by robert on 2008-12-05 11:40:35 +0000 (Fri, 05 Dec 2008) Only write extra bits if there are internal nodes (not sure this is necessary, but shouldn't do any harm.) Extra bits now associated with the point before. Some cosmetics, and first steps toward writing NOD 3.

1 0