Closed turbokadi closed 5 years ago
@theomarzaki Can you look up what I've made. It's just that I thought the significant was 6 but actually it's 7 digits so my coordinates was always out of the map (index -1 in the vector.at()). I've tested with the v2x and it seems to work.
Other than the changes, looks great to me.
Although since the v2x has been updated, the branch cannot really be tested until I merge the other branch. Which I will do soon. So if Fred is happy with the changes, I would like to merge.
No description provided.
:(
It could be great to indicate how to test the feature with a minimum of indications (as "use this Injector test Suite and check into the logs that Y).
I'm going to review the code, test as I can on my pc and quiclkly check the memory
@tigroo31 , perfect.
I will check the accuracy of the waypoints compared to the hardcoded value and the mapper value or I can just check if the mapper values are close to the hardcoded values. Once the branch is merged however, as the sub request here is still circle.
`__GI_raise 0x00007f8bd26aae97 __GI_abort 0x00007f8bd26ac801
int TIME_VARIANT = 0.035;
in to/CreateTrajectory.cpp : Narrowing conversion from 'double' to 'int'
use
const auto TIME_VARIANT{0.035};
auto getClosestFollowingandPreceedingCars(const std::shared_ptr<RoadUser>& merging_car, const std::vector<std::shared_ptr<RoadUser>>& close_by) {
instead of
auto getClosestFollowingandPreceedingCars(const std::shared_ptr<RoadUser> merging_car, std::vector<std::shared_ptr<RoadUser>> close_by) {
The time variant is the stupidest mistake I have made. I fixed it on the other pull request. was actually one of the factors the waypoints weren't working as expected.
for the second function, it is needed, since we already passing the &ref from the for loop. adding && would result in unexpected behaviour
@tigroo31 How can I ignore the coordinates since I must respond something. Maybe I can catch the exception lower in the code but I need to give something to the setter. And normally with the future GPS scope it wouldn't be possible anymore. Maybe I can do an optional but after that, what am I gonna give to the setter.
@tigroo31 I find no leaks other than libtorch. I use valgrind --leak-check=full --show-leak-kinds=all -v ./exchange
std::optional
Le 17 mai 2019 à 16:10, Cactus Mamelut notifications@github.com a écrit :
@tigroo31 How can I ignore the coordinates since I must respond something. Maybe I can catch the exception lower in the code but I need to give something to the setter. And normally with the future GPS scope it wouldn't be possible anymore.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.
@tigroo31 Yeah but we need to have coordinates for the car because it's for the RL, you can't just say your equal to nullopt and Voila. I will be forced to modify all the code from Omar, rather than just drop that request in the GPS frame. Previously the program crash because it attempt to reach a -1 index into a vector without any message, and now it's an intended crash with a more explicit message.
Run the TO and test with the V2x. You can also send some OBUs with netcat for example.
Send that like 3 times, it will generate a recommendation that involve the mapper because there is only one OBU. Check in logs