[03:05:23] Which one causes the gateway timeout? I don't see it. (So, yeah, might have been cache) Do you still see it now? [05:10:26] Earlier it was on the three that still show red fail crosses. But now the text says they pass. So I agree it was probably a cache issue: : https://tools-static.wmflabs.org/bridgebot/bd374ae4/file_62389.jpg [06:49:01] No, I agree with Toby that the results are as expected now, thank you. (re @vrandecic: Which one causes the gateway timeout? I don't see it. (So, yeah, might have been cache) Do you still see it now?) [07:18:38] Is this related to "richer metadata". I'm seeing it from 6 hours ago at Z17132. When I first saw it I thought it was a bug and reported it at T369173. But maybe it is/was deliberate? : https://tools-static.wmflabs.org/bridgebot/576df36c/file_62390.jpg [07:36:51] Yeah, I don’t think we should be seeing that now. I don’t see it if I run the tests in edit mode 🤷‍♂️ (re @Toby: Is this related to "richer metadata". I'm seeing it from 6 hours ago at Z17132. When I first saw it I thought it was a bug and r...) [10:55:21] Now four hours later it still thinks this evaluation took place 6 hours ago... so maybe that's the error : https://tools-static.wmflabs.org/bridgebot/c8799441/file_62399.jpg [11:22:58] Here's a different kind of unexpected error at Z14031. Out of bounds memory access! : https://tools-static.wmflabs.org/bridgebot/c8b4b1fc/file_62400.jpg [11:27:56] I had one of those on May 31st. Response was: “ Fancy, that's a WASM segfault” 🤷‍♂️ (re @Toby: Here's a different kind of unexpected error at Z14031. Out of bounds memory access!) [11:30:40] [here] (re @Al: Never seen this before: [11:30:41] Actual result: { "Z1K1": "Z5", "Z5K1": "Z564", "Z5K2": { "Z1K1": { "Z1K1": "Z7", "Z7K1": "Z885", "Z885K1...) [11:49:11] It’s now working for n=7 but the same failure occurs at n=29 (re @Toby: Here's a different kind of unexpected error at Z14031. Out of bounds memory access!) [12:00:35] Whoop! I just went to the underlying function, made a new test, and got a "Service Unavailable" now at Z17612 : https://tools-static.wmflabs.org/bridgebot/01569f43/file_62401.jpg [12:05:22] …and Z14761 is “out of bounds” for that calculation 🤔 (re @Toby: Whoop! I just went to the underlying function, made a new test, and got a "Service Unavailable" now at Z17612) [12:06:35] Yes, that's where I got to too. It looks like it's coming from Z14733, which could well have an array bound error in the recursion, although I haven't spotted it yet. [12:06:51] "[2024-07-10 12:03:15.550] [error] execution failed: out of bounds memory access, Code: 0x88 [2024-07-10 12:03:15.550] [error] When executing function name: "_start"" (re @Al: …and Z14761 is “out of bounds” for that calculation 🤔) [12:11:17] Looks like error propagation: there’s an evaluation timeout (9000 ms) in the Python, becoming (?) “out of bounds” one level up, and “Service Unavailable” at the level above that 🤔🤷‍♂️ (re @Toby: Yes, that's where I got to too. It looks like it's coming from Z14733, which could well have an array bound error in the recursi...) [18:52:25] 9151 [21:12:18] Definitely! [21:12:19] That looks like a really inefficient way to implement that ackermann like function. Instead of just recursion, it's possible to use bit shifts to speed it up. [21:12:20] The implementation wouldn't be as simple, but maybe a worthwhile trade off in this case? [21:12:21] I'm not sure where I saw the faster alternative implementation though (re @Toby: Yes, that's where I got to too. It looks like it's coming from Z14733, which could well have an array bound error in the recursi...)