!AruGtDGEakIrccthIK:matrix.org

RoboJackets/robocup-software

67 Members
RoboJackets/robocup-software development chat13 Servers

Load older messages


Timestamp Message
24 Dec 2019
02:18:32@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 2M5FLBQB or anyone else who has an opinion. I want the capability to make tests that either use the currently enabled plays (my first thought), or load a playbook file and use those plays (might be a little more robust for sequential tests if each one enables and disables plays). I'm thinking that I should just either modify the base test with an option parameter or make a sub-class that gets the currently enabled plays from the play registry in main and adds them to the enabled plays list in the constructor? Because a lot of the time I want to test the competition playbook as a whole and I don't want to have to modify each of my tests each time the playbook is changed. Is there an easier way to do that?
02:31:33@_slack_robojackets_U2M5FLBQB:matrix.orgchachmu That is what the test system currently does
02:32:37@_slack_robojackets_U2M5FLBQB:matrix.orgchachmu Well... I guess you would have to copy paste the pbk file into the test right now. I guess we could add a way of specifying the location of a pbk instead, wouldn't be very hard
02:38:07@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 So currently it just uses the currently enabled plays? Than what does the self.play_list do? Does it only change the enabled plays if that list isn't empty?
02:38:58@_slack_robojackets_U2M5FLBQB:matrix.orgchachmu Yeah ok I should have read your entire post more carefully. Just have the function that loads the test check if the play_list is actually a path to a pbk. Or you could add another member variable that specifies the pbk and if it's not none have the function use it instead of play_list
27 Dec 2019
10:03:11@antoine:hackerspaces.beantoine joined the room.
30 Dec 2019
20:04:19@_slack_robojackets_U1T86AW4B:matrix.orgguyfleeman Just FYI there's some charter on the league discord about how to log behavior trees for replay. Y'all should chime in
2 Jan 2020
23:58:58@thor:tchncs.dethor joined the room.
6 Jan 2020
19:04:26@justus:matrix.allmende.iojustus joined the room.
7 Jan 2020
17:23:28@diniyar:sibnsk.netdiniyar joined the room.
17:27:08@louis:perthchat.orglouis joined the room.
10 Jan 2020
23:29:49@niels:feneas.orgniels joined the room.
15 Jan 2020
00:37:57@_slack_robojackets_U2M5FLBQB:matrix.orgchachmu Somebody messed up and broke make run for soccer. When you click on the soccer field it immediately seg faults on staging.
00:38:07@_slack_robojackets_U2M5FLBQB:matrix.orgchachmu On further introspection, it might be because of me. I'll have to look into this
00:41:54@_slack_robojackets_U2ZDM7T16:matrix.orgjyc5776 I'm not getting this
16 Jan 2020
18:56:35@_slack_robojackets_UCPA3EW7Q:matrix.orgkwstach Is anyone on the plays side working on/planning to work on the ball placement challenge? IIRC that was planned as a technical challenge for division B.
19:01:50@_slack_robojackets_UCREN25DK:matrix.orghgynai So I was looking at it before, I know Woodward made the play
19:02:03@_slack_robojackets_UCREN25DK:matrix.orghgynai I can take a deeper look into it though
19:11:10@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 I just modified the placement play to actually function and it looks like dribble is the current failure point. Every time we tap the ball we lose it and the dribble has to recapture. It's calls to gamestate are borked probably due to the refactor, shouldn't be too hard to fix. Yeah the dribble behavior is uhh not great. To say nothing of how it will perform irl with our physical dribblers.
19:15:30@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 It's a capture and dribble to point pretty much. It's just that both capture and dribble are bad even in sim, let alone on hardware.
19:16:39@_slack_robojackets_UCREN25DK:matrix.orghgynai I mean when I ran 2M5FLBQB's play I remember it working fine, but that's obviously just in sim
19:18:01@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 It definitely does not work now on staging. After gamestate fixes and hardcoring the point I get dribble bumping/failure.
19:29:50@_slack_robojackets_U2M5FLBQB:matrix.orgchachmu It was tuned like 4.5 years ago so I don't know what you expected to happen
19:31:23@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 It's better than I expected tbh considering we've never had a great reason to do ball placement and are bad at many other more important things.
19:33:03@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 It's probably not very practical to work on until it can be tested on a real robot with relatively decent motion control imo. As that's going to be the key pieces of information.
19:42:45@_slack_robojackets_UCREN25DK:matrix.orghgynai By the way we figured out that Jason's capture and stuff wasn't working because his friction was way too low on grsim
19:43:27@_slack_robojackets_U73HQSDEG:matrix.orgclindbeck3 Oh, I just assumed placement in restarts was it? Is that not it?
19:44:01@_slack_robojackets_UCREN25DK:matrix.orghgynai What do you mean, sorry?
19:57:04@_slack_robojackets_U76UNMBD1:matrix.orgtjones320 Also dribbler is about to change too
19:58:03@_slack_robojackets_U76UNMBD1:matrix.orgtjones320 We should test after we get the new dribblers up and running

There are no newer messages yet.


Back to Room List