Commit: r1771: Code spring clean.
Version 1771 was commited by steve on 2011-01-07 13:27:25 +0000 (Fri, 07 Jan 2011) Code spring clean.
On 07/01/11 13:30, Marko Mäkelä wrote:
Does this tidy up the address search? :-)
Ha! It shouldn't but you never know... But on the subject, has anyone been trying out the index branch? I have been trying to improve the index, but have not been able to do anything about street search on the device. One thing that should be improved however is the sort order, so accented letters are sorted along with their un-accented counterparts (and you should be able to supply a file that specifies the sort order in any way you wish). There are numerous other fixes, and although I've not noticed any improvements, they may have made improvements in ways that I've missed. ..Steve
On 07.01.2011 15:54, Steve Ratcliffe wrote:
On 07/01/11 13:30, Marko Mäkelä wrote:
Does this tidy up the address search? :-) Ha! It shouldn't but you never know...
But on the subject, has anyone been trying out the index branch? I have been trying to improve the index, but have not been able to do anything about street search on the device. Didn't even know this branch exists....
On the other hand, I do get some crashes after the Code spring clean inside the address search in Mapsource, when before search worked without any crashes in Mapsource (used rev 1760 to recreate mdr, and no crash then).
On 07.01.2011 16:01, Steve Ratcliffe wrote:
On the other hand, I do get some crashes after the Code spring clean inside the address search in Mapsource, when before search worked OK I had better look at that, there wasn't supposed to be any change in functionality in that change.
Well I think it will be hard to find the problems. I could not find out any pattern bout the crashes (well the only one being, the further away the searched street is located, the bigger the chance for a crash (Mapsource error reporting popping up, street not found, but Mapsource itself not crashing - with old version in the same case however street found).
On 07/01/11 15:13, Felix Hartmann wrote:
On 07.01.2011 16:01, Steve Ratcliffe wrote:
On the other hand, I do get some crashes after the Code spring clean inside the address search in Mapsource, when before search worked OK I had better look at that, there wasn't supposed to be any change in functionality in that change.
Well I think it will be hard to find the problems. I could not find out any pattern bout the crashes (well the only one being, the further away
OK, I shall back out the two changes that could possibly have an effect on the output. However I don't see how they would affect the index, and in my tests they do not. So if you still have problems could you put up the two mdr files somewhere where I can compare them? Thanks ..Steve
Does this mean I can removed the road-name-pois switch and the cluttered road names? On Fri, Jan 7, 2011 at 10:54 PM, Steve Ratcliffe <steve@parabola.me.uk> wrote:
anything about street search on the device.
-- cheers, maning ------------------------------------------------------ "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ ------------------------------------------------------
On 07.01.2011 16:16, maning sambale wrote:
Does this mean I can removed the road-name-pois switch and the cluttered road names?
On Fri, Jan 7, 2011 at 10:54 PM, Steve Ratcliffe<steve@parabola.me.uk> wrote:
anything about street search on the device. Well road-name-pois should be removed from mkgmap alltogether IMHO. It overloads the POI search, and causes crashes on most gps - even when the map that has been built with road-name-pois is deactivated....
And no, only old generation Garmin GPS work with the address search right now (and to search in other tiles than the active, workarounds are needed). In the end as cpreview that generates a fully working address search for cgpsmapper produced maps is open source now, it should be possible to get mkgmap write correct address search too. I have not succeeded in using cpreview to write working mdr for mkgmap created maps though.
I agree, a good workaround in the past, but should replaced eventually.
Well road-name-pois should be removed from mkgmap alltogether IMHO. It overloads the POI search, and causes crashes on most gps - even when the map that has been built with road-name-pois is deactivated....
And no, only old generation Garmin GPS work with the address search right now (and to search in other tiles than the active, workarounds are
Hope we can get this into mkgmap soon. :)
needed). In the end as cpreview that generates a fully working address search for cgpsmapper produced maps is open source now, it should be possible to get mkgmap write correct address search too. I have not succeeded in using cpreview to write working mdr for mkgmap created maps though.
In trying to understand Garmin's search function, am I right to assume that search works on different features of the map? My observation is that garmin has 3 separate search index (POIs, address/street, and City). POI search is OK but I have observed problems with city search.
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
-- cheers, maning ------------------------------------------------------ "Freedom is still the most radical idea of all" -N.Branden wiki: http://esambale.wikispaces.com/ blog: http://epsg4253.wordpress.com/ ------------------------------------------------------
On 07.01.2011 16:27, maning sambale wrote:
I agree, a good workaround in the past, but should replaced eventually.
Well road-name-pois should be removed from mkgmap alltogether IMHO. It overloads the POI search, and causes crashes on most gps - even when the map that has been built with road-name-pois is deactivated....
And no, only old generation Garmin GPS work with the address search right now (and to search in other tiles than the active, workarounds are Hope we can get this into mkgmap soon. :) needed). In the end as cpreview that generates a fully working address search for cgpsmapper produced maps is open source now, it should be possible to get mkgmap write correct address search too. I have not succeeded in using cpreview to write working mdr for mkgmap created maps though. In trying to understand Garmin's search function, am I right to assume that search works on different features of the map? My observation is that garmin has 3 separate search index (POIs, address/street, and City).
POI search is OK but I have observed problems with city search. throught out road-name-pois from all loaded maps, and i bet that city search works correctly afterwards! (well I'm not sure which city categories are searchable, some are not showing up in search - but that's a style-file problem). (except the usual mkgmap problem of messing up Umlauts and their like in the search function - so best don't use --latin1 or another codepage, then one can search cities with Umlauts too).
Hi
(except the usual mkgmap problem of messing up Umlauts and their like in the search function - so best don't use --latin1 or another codepage, then one can search cities with Umlauts too).
The index branch has better treatment of accents in the mdr file. So I would be interested to hear your comments on it. All sorting within the img tiles is still done the old way so there may still be some problems, but it would be interesting to know what now works and what still doesn't. ..Steve
In the end as cpreview that generates a fully working address search for cgpsmapper produced maps is open source now, it should be possible to get mkgmap write correct address search too. I have not succeeded in using cpreview to write working mdr for mkgmap created maps though.
I don't know cpreview as an user. I know that it is open source. Which input files does cpreview expect? The same as mkgmap (a compiled map) or other ones? If they are the same, it should be possible to generate indexes with both tools and compare the output for differences. But I think it would be safer to look into the sources. Rgards, JOhann
participants (6)
-
Felix Hartmann -
Johann Gail -
maning sambale -
Marko Mäkelä -
Steve Ratcliffe -
svn commit