berestovskyy wrote:

MdSyah wrote:Would you please put back the highlight of dead end u turn check back for Malaysia like those of the UK.

So, maybe it's easier to turn it back on for all countries but US then?

No, not yet anyway, as this goes against staff's express wishes. So far I believe the UK is the only country to get any sort of special permission from staff to leave them off.

typeRank
15 freeway
14 major highway
13 minor highway
12 ramp
11 primary street
10 street
9 service road
8 dirt road
7 parking lot road
6 private road
5 walking trail
4 pedestrian bdwk
3 stairway
2 railroad
1 runway

type
1 street
2 primary street
3 freeway
4 ramp
5 walking trail
6 major hwy
7 minor hwy
8 dirt
9
10 ped boardwalk
11
12
13
14
15
16 stairway
17 private road
18 railroad
19 runway
20 parking lot road
21 service road

I don't know about the new Ferry type, I haven't tested it yet.

Taco909 wrote:Okay, I'm not sure if this belongs in Validator or if it is a WME bug.

With the week's changes, my first custom highlight is now nul elevation (and when I clear it, it comes back on next reload)
That's fine, I can live with it... but it seems that for the most part, I only see it after I've edited a segment... such as after deleting a connected segment or unneeded junction node.
For the unneeded junction node I would expect the purple highlight to take precedence, but I'm seeing others that were not highlighted before a save and pop up green after saving.
In one case, after a save, a street two segments removed from any editing popped up nul.

At first I thought I still had highlighter set to display recent edits, but confirmed that it was not set.

They are legitimately nul elevation, but I have not yet been able to confirm whether or not they were nul before the edit-save operation.

The only reason you started to see null elevation is that JNF wasn't working properly for a while, and JNF suppresses null elevation (shows 0 instead). So with JNF suppressing null elevation, Validator didn't see it, and couldn't show it.

Taco909 wrote: Please be aware that some apparent overlaps and spurious geonodes were placed intentionally for TTS wayfinding assistance.
It is not uncommon for a junction to have a 90 degree turn, but the geonode is placed so closely that it appears to be a shallower angle. These are intentional to force a "Turn ____" instruction rather than "Stay to the ____"

Likewise, there may be some cases where it is desired to execute the turn instruction early, but to overlay the departing segment onto the feeder segment for a short distance.

Apparent spurious geonodes, yes, but 90° turn-instruction micro-doglegs wouldn't show an "overlap". The overlap highlight only shows when the segments run on top of each other for a moment, so making such a dogleg would remove that highlight, unless of course the segment overlapped the other segment later on.

As for the early instructions, IMO it's better as an editor to lay the departing segment just next to, but not directly coincident with, the continuing segment.

In short, IMO the overlapping check is still quite useful to detect improperly-set "dogleg" constructions (usually where the geonode was accidentally snapped to the other segment) and to let you know to give the latter scenario just a little bit of leeway.

PesachZ wrote: As far as the highlight, my experience it only checks for two segments connected to the same junction, at the same turn angle . So am apparently overlapping segment with a dogleg wouldn't get a highlight, provided the dogleg caused the two segments to have different turn angles.

Definitely not so – I see it often with very low but nonzero angles, for example in ramps with long acceleration/deceleration lanes where the last geometry handle is pretty far from the junction node but pretty close to the other segment (on the other axis). This comes up mostly when I'm editing freeways.

CBenson wrote: I would go further however as In my personal opinion angles of less than about 20 degrees should be avoided whenever possible. My reasoning is that the road closure feature in the app is next to useless for intersections with small angles. But there is the question of whether we should be mapping to the make the road closure feature work or whether the road closure feature should be made to work with the map that conforms to reality.

You bring up exactly my concern. I will not edit the map to enable a feature that doesn't really work anyway. That feature needs a lot of work, and it's not up to us to fix it.

Thanks berestovskyy!!

Validator is doing something real strange in the beta editor when I try to create a new segment and when I try to edit a segment name.

I made a little video to show you:

Note the URL after I try to edit segment details. I narrowed it down to Validator. It seems to work fine in the normal editor.

Ay, sorry! Thanks! I don't normally use that option anyway

berestovskyy wrote: 3. To use a backslash inside a string you need to escape it, i.e. to match an alphanumeric character use "\\w" and to match a dot use "\\."

Why do you need two backslashes? I have always used a single backslash to escape (like \. to match a period and \d to match a digit and it's worked. But I notice that Waze also uses the \\ double backslash in their TTS regex.

berestovskyy wrote:

sketch wrote:Why do you need two backslashes?

You don't if you put your regexp into an edit field, i.e. when you use Validator interface for the custom checks you should use just \w, \d and \.

But you do need to double each backslash when you put the regexp into a JavaScript string, i.e.

Code: Select all

"regexp": "/this is a JS string, so we use \\w and \\. here/",

In JS, Java, C, C++ and many other programming languages we use a backslash as an escape character. Have a look at the Using special characters in strings at MDN.

Ah, I see. So I have to escape once for JavaScript and then escape again for regex. Thanks!

It is a nice idea. But let's recall that Validator currently only checks segments, and I don't think there are any plans to expand past that at this time.

Agreed, a "service road" check would be nice. According to staff, the entire type is going to be removed, but then again they've been saying that for two and a half years...