Notices

Yes, I met local lib. I will have to look into which libs are pulled there and why plain boost is used. Thanks for noticing!

For those who observed issues with boost - please check if http://repo.merproject.org/obs/home:...la.armv7hl.rpm works. I bundled the missing library and I hope it is sufficient to fix the problem

PS: Already got the feedback - all should be fine now.

OSM Scout Server has been packaged into Flatpak - available at https://github.com/rinigus/osmscout-...server.flatpak . This allows you to use the server on desktop easily and develop against it. For example, modRana and Pure Maps.

I'll submit the server to Flathub, just may take some time. The server will have to be adjusted on desktop further - please ignore messages regarding missing icons. But, it is already now usable for testing clients against it

The new release was done on desktop only - there are no relevant SFOS changes.

Indeed! So far I have been always testing OSM Scout Server related functionality on my Sailfish OS device, but this way I can work on it directly on desktop as I do for basically everything else.

All in all, it works very well - I've tested map data download, tile rendering and place search and all seem to work fine! Actually, the Mapnik generated tiles seem to look better & have more details on desktop than when rednered on Sailfish OS. I guess this is due to some libraries being available that can't be used on Sailfish OS or possibly reduced resource constraints ?

BTW, I've noticed the socket activation toggle is still there in the flatpak version. IIRC you have mentioned socket activation is currently not possible with flatpak, so I guess the option should be hidden when running from flatpak ?

Thanks a lot for this! Having OSM Scout Server so easily available on desktop is really really nice & I think I'm certainly not the only one who will find it useful!

yep, its a bug. I'll have to fix it

OSM Scout Server is now available at https://Flathub.org . So, its compiled for all supported platforms (32- and 64bit variants of x86 and ARM). Whether it works, that's another question.

I'll try to keep all releases for SFOS and Flathub in sync.

New maps are out. Sorry for delay with the import - had an issue with intermediate database and had to fix it.

New release is out 1.13.0

Its relatively small in terms of changes - decided to release them to get it published. Search results now include postal code as a name, if the name is missing. Should be handy for those who use postal codes as a main mean of navigation.

The server has added support for Qt 5.11. For that, I had to change SQL backend used by Mapbox GL as a part of working around https://bugreports.qt.io/browse/QTBUG-39957 .

There were updates in documentation, Spanish User's manual in particular (thank you @carmenfdezb!).

Edit: Let me know if it doesn't work for 2.x SFOS series - we can always add a target at OBS

Tested on a Jolla 1 under SFOS 2.2.1.18:

1. PureMaps displayed no map after updating OsmScoutServer.
2. OsmScoutServer complained about missing the "mapboxgl world overview map" and hence did not start.
   I could swear I had downloaded it before (and OsmScoutServer was definitely working before the update). The MapManager also showed it as "missing". When checking for updates there, other maps were determined as updatable.
   So I just downloaded the "missing" GL world overview map.

Besides this little prelude, everything seems to be working as before, when walking around the block using voice navigation.
Are there specific things to look at?

P.S.: When checking that the "MapboxGL world overview map" has been downloaded correctly (sorry, missed to do that before the download), I noticed that mapboxgl/packages/tiles-world.sqlite exactly matches the size reported by OSMScoutServer on startup (13520896 bytes), but mapboxgl/glyphs/glyphs.sqlite does not: 75730944 bytes reported by ls -l, in contrast to ...
INFO: Missing data: Mapbox GL fonts (75,730,947)
INFO: Missing data: Mapbox GL World overlay (13,520,896)
I have not checked any other sizes, as they need to be accumulated over various files in multiple directories.
Well, if just the reported sizes are incorrect (either by OSMScoutServer or ls), that doesn't harm (in contrast to if incorrect file sizes are used internally in OSMScoutServer).

@olf: I was mainly concerned with whether it will start at all due to the building performed using the latest OBS release (SFOS 3).

RE glyphs: no idea why the file size in the database is different. Its as if the file shrank after compression/decompression by 3 bytes. I am writing down sizes using "du -shb" which should give the correct size while packing. Let's keep it as a small mystery, as long as it doesn't disturb function.