Parsing OSM tiles on WatchOS fails due to feature id overflow

[JeffLutzenberger]

It looks like integer feature ids should be explicitly cast to UInt64 on WatchOS. I have a branch with tests for this.

Gist for discussion: https://gist.github.com/JeffLutzenberger/369d3062dafb5925d9641119eef41edd

(MVT tile spec: mapbox/vector-tile-spec#94)

[trasch]

@JeffLutzenberger Thank you for bringing this up.

There are two conflicting definitions here:

GeoJSON:
o If a Feature has a commonly used identifier, that identifier
SHOULD be included as a member of the Feature object with the name
"id", and the value of this member is either a JSON string or
number.

Meaning, a GeoJSON id could be any Int, positive or negative.

MVT Spec:
var id: UInt64 {...}

So an Unsigned Integer in vector tiles...

In the end, we need a solution that works with both:

• GeoJSON FeatureId needs to support Int and UInt64
• The MVT decoder needs a change

Do you have an example URL for a OSM tile with an overflowing id?

[trasch]

... and I have to admit that it was rather shortsighted of me to not handle this correctly right from the beginning... 🙄

[JeffLutzenberger]

it seems subtle though. I'm not 100% sure, but this might be a result of arm64_32 on the watch. Regardless, maybe just explicitly parsing Int64 would suffice.

I'll see if I can get you a tile, or branch with a tile on disk that demonstrates the issue

[JeffLutzenberger]

Here's one that fails. Parse with Ultra 2 simulator or device:

osm-8-41-101.pbf.zip

[trasch]

@JeffLutzenberger

I think it should work now with Outdooractive/mvt-tools#18 and #46, but I will test it later

[trasch]

@JeffLutzenberger The new uint64Value property on Feature.Identifier should give you consistent access to a UInt64 id for all OSM features, on all platforms. I hope this solves your problem... (?)

[JeffLutzenberger]

Testing now. Looks good so far!

[trasch]

@JeffLutzenberger How is it going? Everything good so far?

[JeffLutzenberger]

So far, so good! Thanks for the quick response and engagement on this.

[trasch]

I will merge this next week if everything is working now, so that this can be closed... :-)