[01:43:18] https://tools-static.wmflabs.org/bridgebot/9b30bd40/file_66391.jpg [01:43:49] Like this section could be replaced with Wikifunctions, making it more interactive and hands-on [01:44:24] If we want the world to have access to the sum of human knowledge we should make the knowledge easy to comprehend and understand [02:14:53] I tried to add a test to Z18784 that would do the same thing: Z19660. But I probably did it all wrong. Can anyone maybe help me to fix it? (re @Feeglgeef: You set the input to a function call, make that call Z16568, input "Natural Language", input what language you want, input Russi...) [02:17:58] https://tools-static.wmflabs.org/bridgebot/69956728/animation_gif.mp4 [02:18:00] https://tools-static.wmflabs.org/bridgebot/69956728/animation_gif.mp4 [02:18:02] Fuck [02:18:03] Sorry [02:18:21] What exactly are we trying to do here (re @amire80: I tried to add a test to Z18784 that would do the same thing: Z19660. But I probably did it all wrong. Can anyone maybe help me ...) [02:18:30] I didn't mean to send this (re @Feeglgeef: ) [02:21:50] All the other tests for Z18784 get a string as an input. I tried to add a test that gets this string from the language name that is defined in a language object. [02:22:31] So you want to grab the Russian label for Duala? (re @amire80: All the other tests for Z18784 get a string as an input. I tried to add a test that gets this string from the language name that...) [02:30:01] Yes. [02:30:18] Feel free to edit the same test. [02:33:05] It should be fixed [02:33:11] Idk why it's not passing [02:33:28] https://tools-static.wmflabs.org/bridgebot/e65e7230/file_66394.jpg [02:33:34] This is what it should look like [02:40:28] Thanks... Initially I added an intentionally wrong result, because I don't trust tests that I've never seen failing. [02:40:40] And it was not connected, and I connected it. [02:40:56] And I fixed the expected result to the actual one. [02:41:00] But it still fails. [02:41:06] Yeah [02:41:23] It's probably a problem with Wikifunctions? [02:42:01] The same exact code is performing completely differently in the Try vs in the test [02:43:15] And the "Details" of the failure are not particularly helpful. [02:43:19] https://tools-static.wmflabs.org/bridgebot/c2af2143/file_66395.jpg [02:43:35] That's rough (re @amire80: ) [02:43:36] Sorry, let me send it in English. [02:44:27] Lemme try forcing it to be what we want with source edit [02:45:24] ``Actually no, I'm not sorry, it is by design that it's in Russian.`` [02:47:08] https://tools-static.wmflabs.org/bridgebot/c85f0bf0/file_66396.jpg [02:47:32] I can't figure out the issue [02:47:36] It looks right [02:54:42] (What most people don't see in this screenshot is that there is a mistake in the translation of one of the UI strings into Russian. I fixed the message on translatewiki, and it should be deployed to Wikifunctions next week. I also fixed the documentation for translators in the hope that translators to other languages will know how not to make a similar mistake, and [02:54:43] fixed the tran [02:54:43] slations into other Slavic languages, all of which have a similar grammar issue with plural forms.) (re @amire80: ) [03:36:37] Had this been solved? (re @Toby: I agree that P5186 does not seem to be returned as a statement when I run Z6825 with an argument of L20643. I'm trying to see wh...) [03:45:08] You inspired me to make the renderer useful. I finally got the JS implementation of Z14310 to work, so now it can run in time, and therefore we can switch to using Z14340 for Z14280... which means we can finally see the commas in numerical results: : https://tools-static.wmflabs.org/bridgebot/c95e947f/file_66397.jpg [03:48:00] No. I'm sure we'll be told in the weekly news once it has been. (re @cvictorovich: Had this been solved?) [03:51:26] Yes, "Search for Type to select" might explain it better. I like having the suggested types already there. (re @amire80: And the placeholder says "Select type". I understand the word "Select" as selection out of several presented options and not so ...) [03:55:32] I think Integer (or natural number) and Typed list should be added (re @Toby: Yes, "Search for Type to select" might explain it better. I like having the suggested types already there.) [03:55:57] But there should be suggested types, for sure [04:00:04] Like this, though, it does make it a little big : https://tools-static.wmflabs.org/bridgebot/e324929a/file_66398.jpg [04:01:49] Or like this, to save vertical space, given that type is already implied by the dropdown and the options : https://tools-static.wmflabs.org/bridgebot/3854f37a/file_66399.jpg [04:06:06] Perhaps Lexeme could be added to this? Otherwise it LGTM [06:16:09] Yeah, that’s this bug *T366246* (re @amire80: I tried to add a test to Z18784 that would do the same thing: Z19660. But I probably did it all wrong. Can anyone maybe help me ...) [06:23:48] I tried to make it more intuitive, but *T366459* (re @amire80: Thank you so much for this. Few things are more satisfying than feeling than something is very odd and then hearing someone else...) [07:29:03] That's great! I was planning to do something like this myself, but you were first, which I really appreciate! Thanks. I will take a look at your work soon, and see whether to switch the renderer. (re @Toby: You inspired me to make the renderer useful. I finally got the JS implementation of Z14310 to work, so now it can run in time, a...) [08:38:33] I see, it is already switched! 😄 That's awesome [08:54:55] There can be quite a delay but the cacheing is working nicely. I wonder how many numbers would get cached over time? (re @vrandecic: I see, it is already switched! 😄 That's awesome) [09:06:51] The same [09:50:30] Z19676 is a little stress test. I was getting some red errors switching between short and long forms of the result validation but they are not occurring now. (re @vrandecic: The same) [09:53:28] Like this : https://tools-static.wmflabs.org/bridgebot/d687c0e7/file_66405.jpg [10:35:18] Yes, I just tried switching to German, and it was all red at first, but after a few reloads it sorted out. (re @Al: Z19676 is a little stress test. I was getting some red errors switching between short and long forms of the result validation bu...) [10:49:40] PS I expect that the speed would still benefit significantly from having an inbuilt implementation of the "apply" function. T357858 (re @vrandecic: I see, it is already switched! 😄 That's awesome) [10:58:54] If you do visit Z873 @vrandecic , please can you connect all the tests, they're nice ones for a hard-to-understand function. (re @wikilinksbot: Z19676 – primorials for prime numbers 11 to 47) [11:08:39] I tweaked Z13048. (re @Toby: PS I expect that the speed would still benefit significantly from having an inbuilt implementation of the "apply" function. T357...) [11:15:45] Now it's identical to Z14353. Denny showed this improvement when he featured it as FOTW. (re @Al: I tweaked Z13048.) [11:17:35] Yeah, I just noticed that 🙄 (re @Toby: Now it's identical to Z14353. Denny showed this improvement when he featured it as FOTW.) [11:27:38] Reverted 😎 (re @Toby: Now it's identical to Z14353. Denny showed this improvement when he featured it as FOTW.) [11:32:48] In the message `wikilambda-functioncall-metadata-errors` in the WikiLambda extension, can $1 ever be 0? [12:08:52] At least it pushed the faster implementation to the top https://www.wikifunctions.org/w/index.php?title=Z13036&curid=23026&diff=137122&oldid=130685 [12:08:52] It is actually the selection of the function that is the bigger concern. I’m guessing we end up passing function objects into your JavaScript when it really only needs references to those objects. But that brings us back to *T366459*… (re @Toby: Now it's identical to Z14353. Denny showed this improvement when he featured it as FOTW.) [12:10:34] …or *T366403*, even. [14:16:47] Perhaps you could use a local override for now (re @jhsoby: The Wikifunctions favicon still looks terrible with a dark background. Any plans to fix it?) [14:24:30] Like here the number could be an input and the Cipher letters and output would autogenerate (re @Feeglgeef: ) [14:25:02] It would appear that there are no current plans: T326094 (re @jhsoby: The Wikifunctions favicon still looks terrible with a dark background. Any plans to fix it?) [15:54:11] Is this a problem? (re @Al: It would appear that there are no current plans: T326094) [16:09:28] done (re @Toby: If you do visit Z873 @vrandecic , please can you connect all the tests? They're nice ones for a hard-to-understand function.) [16:19:06] Not for me, but that is the ticket for anyone who is interested. (re @cvictorovich: Is this a problem?) [16:24:14] Not for me either (re @Al: Not for me, but that is the ticket for anyone who is interested.) [21:58:31] Isn't the selection of the function highly cacheable? Once an English speaker runs any natural number function, the comma function will always be the answer to "choose the English display function for numbers". Maybe that answer then just gets looked up every time? (re @Al: At least it pushed the faster implementation to the top [21:58:31] https://www.wikifunctions.org/w/index.php?title=Z1 [21:58:31] 3036&curid=23026&diff=...) [22:12:33] 🤔 I think so, but it doesn’t seem to get quicker after the first number, only when using the same number again 🤷‍♂️ (re @Toby: Isn't the selection of the function highly cacheable? Once an English speaker runs any natural number function, the comma functi...) [22:19:53] True. I don't feel it either. I'd be interested if the team can tell us whether this kind of caching is operational. I guess in some other functions it would be rare to ever ask to run the exact same arguments. So maybe it should be turned on for only some functions. (re @Al: 🤔 I think so, but it doesn’t seem to get quicker after the first number, only when using the [22:19:53] same nu [22:19:54] mber again 🤷‍♂️) [22:33:15] I raised this as a side issue in T363908#9960647 but I note that I did not phrase it as a question 😉 (re @Toby: True. I don't feel it either. I'd be interested if the team can tell us whether this kind of caching is operational. I guess in ...)