Yep. Same thing happened to me after installation of .46 at the service center
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
plumazul
Member
Well, looks like I'm not alone on this. I talked to service yesterday and they were stumped. It was bumped up to the engineers. It appears the API status no longer includes the roof so the app assumes I don't have one.
I wonder how many more this has happened to, but they have not noticed yet?
Last edited:
Do you have the .46 850.. firmware or the .46 3387.. firmware? Looking at EV-FW, not a lot of people have the .46 850 firmware so that may be why more people haven't come forward saying they're having issues.
I called Tesla and informed them of the issue. I mentioned others are reporting the same issue on the forums so its not just my car/app. They're escalating it and promised an update later today.
So I did an API call (thanks to the great video @Akrion) to check out what the car is reporting for its state. It looks like the car is reporting the sunroof is in an "unknown" state. It also looks like something has changed on Tesla's end and they have significantly cut the number of parameters you get back on the API call for vehicle state (including uncorked status). See below for the response I got on my car.
Hopefully this is a bug and they'll fix it soon.
"response": {
"api_version": 3,
"autopark_state": "standby",
"autopark_state_v2": "standby",
"autopark_style": "standard",
"calendar_supported": true,
"car_version": "2017.46.8 50d6145",
"center_display_state": 0,
"df": 0,
"dr": 0,
"ft": 0,
"homelink_nearby": true,
"last_autopark_error": "no_error",
"locked": true,
"notifications_supported": true,
"odometer": 7592.954591,
"parsed_calendar_supported": true,
"pf": 0,
"pr": 0,
"remote_start": false,
"remote_start_supported": true,
"rt": 0,
"sun_roof_percent_open": 0,
"sun_roof_state": "unknown",
"timestamp": 1511643374297,
"valet_mode": false,
"vehicle_name": "Stela"
Yes it has been already noted that Tesla decided to completely remove multiple parameters from their response on the vehicle state API end point. Quite disappointing :/. Its their API so obviously they can do whatever they want but I am still in disbelief that there is no official lib and documentation etc.
Wonder if there will be other calls available to get some of the dropped data? It may make sense to trim back the main status call. No need to constantly report something that is static.
If course, this does not apply to sunroof. It should be reported. I would think windows should be reported as well, even if not possible to remotely change the window state.
If course, this does not apply to sunroof. It should be reported. I would think windows should be reported as well, even if not possible to remotely change the window state.
thegruf
Active Member
fwiw 17.36 => 17.46 noticeably for me
- fixed the very annoying Bluetooth bug introduced with 17.36 with no incoming audio channel when making a call out from the car.
- improved auto dipping headlights immeasurably from utterly uesless to actually (mostly) quite impressive, though they still dont dip in streetlit areas as AP1 did.
- improved AP2 a ton, still not without its foibles of course but breadth of usability much improved and definitely more relaxing. First time I have seen it reliably detect a nearside kerb without a white line. Does have a habit on dual carriageways/divided highways of false detecting a 3rd lane where the median is though! Lane change much improved though where .36 was apparently using the rear facing side camera to detect traffic on the lane you were changing in to and would hesitate before allowing the change, .46 just changes bam. Dont understand why following distances are still different on highways vs local roads.
- Whole car just seems to drive nicer? Some suble work on the drivetrain/motor software?
- Chill mode - tried it once - never again, borderline dangerous during an overtake to find that all the power has gone. Maybe it should have a "kickdown" so if you WOT it wakes up from Chill and turns Ludicrous. If I want the performance of a Leaf I'll buy a Leaf thankyou.
I'm on .44 and my car frequently reports unknown for the sunroof state, and has over the past year's firmware versions. I think that's normal. Sunroof has always worked just fine.
I've heard folks reporting changes/missing info, but I haven't been able to find any details. What's gone? I'm not familiar with the uncork parameter, so unable to check. Someone have any details on these changes?
plumazul
Member
Just for clarification, the problem being reported is with the remote access to roof functions. While using local controls it works fine.
Okay thanks, yeah, remote and local has always worked fine for me. Vent roof shows up and works.
But I see that sun_roof_installed is missing from the vehicle status that Red_DS posted. Guessing that's why the app thinks the car doesn't have a sunroof.
Edit: Full list of what I see missing from Red_DS's vehicle state as compared to .44 -
sun_roof_installed
rhd
has_spoiler
spoiler_type
perf_config
roof_color
third_row_seats
seat_type
rear_seat_heaters
exterior_color
rear_seat_type
car_type
wheel_type
dark_rims
Last edited:
plumazul
Member
Just got a call from my SC confirming that an API change broke the app. They plan to push an App update, NOT an API fix.
Yeah, this is starting to look like API cleanup to me. As other folks have mentioned, a bunch of that stuff is essentially static - I think things like a sunroof being installed is already part of the option codes on the vehicle (RFP2). So somebody probably went through and realized there's a sun_roof_installed vehicle status that's redundant.
SucreTease
Teslarian
thegruf
Active Member
yeah sure I get this, but the distances for each setting onthe control are not the same on local roads as highways, which leads to repeated altering setting.
I freqently have the control set to 4-6 on local roads otherwise it is too close, and if you dont set it down to 1-2 on highways the following distance is so large that you just get cars jumping in front of you.
Maybe it is regional UK thing as I don't see much talk about it.
Bit weird nonetheless.
don123
Member
Sorry to hear about those problems, fellas. Hope they are resolved soon and you actually get to enjoy your cars. No, I'm not Elon's buddy, but I really enjoy my X ... especially AP.
don123
Member
AP and no road paint?
1) This past week AP repeatedly drove smoothly, day and night, at 55 mph on several 1,000 foot straight zones of freshly paved undivided highway with absolutely no paint. It worked with and without other traffic in both directions. It didn't even flinch. Way to go, Tess. (Had hands on the wheel and the adrenalin flowing.)
2) The other night AP repeatedly drove a half a mile smoothly at 30 mph on a curvy paved road with no paint whatsoever, heaps of patches, and a poorly defined right hand edge because of gravel and broken up asphalt.
3) AP even drives on our local straight unpaved (farm) road, but sooner or later heads for the cows ... mostly sooner.
Of course AP excels on reasonably marked highways and local roads ... IMO.
1) This past week AP repeatedly drove smoothly, day and night, at 55 mph on several 1,000 foot straight zones of freshly paved undivided highway with absolutely no paint. It worked with and without other traffic in both directions. It didn't even flinch. Way to go, Tess. (Had hands on the wheel and the adrenalin flowing.)
2) The other night AP repeatedly drove a half a mile smoothly at 30 mph on a curvy paved road with no paint whatsoever, heaps of patches, and a poorly defined right hand edge because of gravel and broken up asphalt.
3) AP even drives on our local straight unpaved (farm) road, but sooner or later heads for the cows ... mostly sooner.
Of course AP excels on reasonably marked highways and local roads ... IMO.
don123
Member
Slow computers?
AP's dive to the side at roller-coaster road crests can be eliminated. Simply set your AP speed much lower as you approach the peak of the crest. The sharper the peak, the slower you should go. Of course you need to have no cars behind you.
It appears that the car's computers simply need more time or computing power. (MX, AP2/HW2, EAP, FSD)
Can someone confirm or debunk that observation?
The good news is that Elon has said he would install, if need be, faster computers in those cars ordered with EAP/FSD. Evidently he meant at no cost to the car owner.
AP's dive to the side at roller-coaster road crests can be eliminated. Simply set your AP speed much lower as you approach the peak of the crest. The sharper the peak, the slower you should go. Of course you need to have no cars behind you.
It appears that the car's computers simply need more time or computing power. (MX, AP2/HW2, EAP, FSD)
Can someone confirm or debunk that observation?
The good news is that Elon has said he would install, if need be, faster computers in those cars ordered with EAP/FSD. Evidently he meant at no cost to the car owner.
Possible alternative... I’ve been watching the IC and it shows the sudden and false left or right turn in road. Let’s concider the cresting case of the center line. Because of the foreshortened perspective it really does look like a right turn. The right line looks like a left turn. If either drops out or one has priority, false turn following is result. Similar for dip.
So I think that depth perception is weak. Possibly in addition to other effects.
SucreTease
Teslarian
The point was that the setting is based upon time, not distance. For example, each increment represents a half-second worth of following distance (I don't know the actual amount of time, as I am awaiting delivery of my Model S). Two seconds of following distance at 70 mph is way more distance than at 35 mph.
