[12:47:34] I'm currently seeing an issue at Z24322, where the three implementations are "Running" indefinitely (or at least for over a minute after refreshing the page). Do others see similar behaviour? [13:00:40] I did… but I think the WikiLambda system heuristic is scratching its head. All those Python failures from *T391435* [13:00:42] perhaps? I’ve asked for someone to have an initial look at that as a matter of priority, but if @Sannita wouldn’t mind checking that the message is getting through, not still waiting for triage… (re @u99of9: I'm currently seeing an issue at Z24322, where the three implementations are "Running" indefinitely (or at least for over a minu...) [13:02:40] I'll ping James to see if anyone can take a look at this asap, but I cannot promise anything (re @Al: I did… but I think the WikiLambda system heuristic is scratching its head. All those Python failures from T391435 [13:02:42] perhaps? I’ve ...) [13:04:49] I don't see why the WikiLambda system heuristic would have a role here. If it hasn't chosen the fastest implementation yet, that's okay, the test should eventually time out. I'm not even seeing a timeout though, they're still "running". (re @Al: I did… but I think the WikiLambda system heuristic is scratching its head. All those Python failures from T391435 [13:04:51] perhaps? I’ve ...) [13:07:45] Thanks. We’re not panicking 🙏 (re @Sannita: I'll ping James to see if anyone can take a look at this asap, but I cannot promise anything) [13:09:59] Yeah, all the tests have completed except for the last one. Shall we try connecting it so we can see whether it’s completing in the implications? [13:11:20] I think leave it disconnected in case it's a new type of bug. I'll file a phab? (re @Al: Yeah, all the tests have completed except for the last one. Shall we try connecting it so we can see whether it’s completing in ...) [13:13:04] Looks like your N is null. (re @u99of9: I think leave it disconnected in case it's a new type of bug. I'll file a phab?) [13:14:11] Ohhh.. that would do it. I must have entered it too fast for the parser to load. Okay, I'll fix and it should be sweet again. (re @Al: Looks like your N is null.) [13:15:09] Could be… but an empty string in Z1358K1 should not be saved. (re @u99of9: Ohhh.. that would do it. I must have entered it too fast for the parser to load. Okay, I'll fix and it should be sweet again.) [13:15:36] Although oddly it appears to block the entire list. Did you say you see all tests completed? : https://tools-static.wmflabs.org/bridgebot/e14b5ebc/file_70202.jpg [13:15:43] Meh, you know what I mean. (re @Al: Could be… but an empty string in Z1358K1 should not be saved.) [13:16:36] Yeah, looking at each implementation… but the disconnected one doesn’t feature there. (re @u99of9: Although oddly it appears to block the entire list. Did you say you see all tests completed?) [13:16:48] I feel like it's common to save blank keys when the parser isn't fast enough. (re @Al: Could be… but an empty string in Z1358K1 should not be saved.) [13:17:49] Not everyone is as speedy as you… but it’s definitely a bug of some description. (re @u99of9: I feel like it's common to save blank keys when the parser isn't fast enough.) [13:18:32] Ah, those are cached from 1-3 hours ago, so they haven't been affected by my new infinity-test just yet. (re @Al: Yeah, looking at each implementation… but the disconnected one doesn’t feature there.) [13:21:37] Okay, I'll try to document this reproducibly sometime. (re @Al: Not everyone is as speedy as you… but it’s definitely a bug of some description.) [13:23:23] Actually, do you think I should fix it now, or let the team figure out if this is causing runaway processes? (re @u99of9: Although oddly it appears to block the entire list. Did you say you see all tests completed?) [13:27:11] It’s probably easier to create an equivalent test case, isn’t it? (re @u99of9: Actually, do you think I should fix it now, or let the team figure out if this is causing runaway processes?) [13:27:49] so let the current one continue to block the lot?? (re @Al: It’s probably easier to create an equivalent test case, isn’t it?) [13:34:40] T392905 (re @u99of9: so let the current one continue to block the lot??) [13:39:16] I don’t understand this part, to be honest. I’ve only seen it occur temporarily before now. Is directly removing it from the function’s list of test-cases a legitimate and expedient option to consider? (re @u99of9: so let the current one continue to block the lot??) [13:42:51] I've also never seen it persist before, but then again I may not have made this exact mistake before. I'm not sure what you mean by "removing" it. It's already not connected, so wouldn't appear in the function's json list of tests. (re @Al: I don’t understand this part, to be honest. I’ve only seen it occur temporarily before now. Is directly removing it from the [13:42:51] fun...) [13:43:36] Ah, yes… good point. (re @u99of9: I've also never seen it persist before, but then again I may not have made this exact mistake before. I'm not sure what you mea...) [13:47:43] I guess it’s picking it up from some internal table, so even pointing the case at some other function might not work. Then you either leave it or delete it… unless you actually fix it. (re @u99of9: I've also never seen it persist before, but then again I may not have made this exact mistake before. I'm not sure what you mea...) [13:54:31] Thank you for filing this! We've observed the same behavior in https://www.wikifunctions.org/wiki/Z19509: the errors caused return an error log so big that the request fails with "This result was truncated because it would otherwise be larger than the limit of 12,582,912 bytes." (re @u99of9: T392905) [13:58:30] Yikes. I can't see any broken test there though. Maybe it's just a coincidence that it happened at the same time as I made a broken test? (re @genocation: Thank you for filing this! We've observed the same behavior in https://www.wikifunctions.org/wiki/Z19509: the errors caused retu...) [14:03:34] I saw this in the logs from the 9th of April, so maybe unrelated to the broken test. Will look into it and update the phab task [14:07:12] Feel free to fix the test to test the theory if it gives you more info. I'll leave it with you as it's past 🎃 time for me. (re @genocation: I saw this in the logs from the 9th of April, so maybe unrelated to the broken test. Will look into it and update the phab task) [14:10:11] Thanks, Toby. Sleep well! (re @u99of9: Feel free to fix the test to test the theory if it gives you more info. I'll leave it with you as it's past 🎃 time for me.)