[00:13:08] Now that we have types which both resolve to integers in python, there are cases like Z16308 where python == fails, I suggest we split Z13052 into one that also first checks type, and one that just follows the python mapping. If splitting is the right thing to do, then which should stay in the current location? [00:25:10] It already has a composition that checks the type first but it’s disconnected because object type equality is currently an unsolved problem 🤷‍♂️ (re @Toby: Now that we have types which both resolve to integers in python, there are cases like Z16308 where python == fails, I suggest we...) [00:31:55] It's also disconnected because it circularly calls itself at the moment. Once there's an untyped version, it can call that. But you're right, the type check will need to work. I see Z16310 currently fails. [00:36:05] You should be able to do it in JavaScript for the time being. Months are integers and Natural numbers are BigInts in JS. (re @Toby: It's also disconnected because it circularly calls itself at the moment. Once there's an untyped version, it can call that. But ...) [00:52:44] When I run a month through Z15969 it looks like it has the info, but we need to check more than just the Z1s? [01:07:21] I think I fixed it. There’s a bug in Z866 that allows it to produce a false positive when the strings are references. It just about functions using object equality instead of string equality, but it comes close to timing out, so you can’t do anything with the inferred equality 🤷‍♂️ (re @Toby: When I run a month through Z15969 it looks like it has the info, but we ne [01:07:21] [01:07:22] ed to check more than just the Z1s?) [01:10:15] It seems to work! I was going down a different rabbit hole, but am happy to have that one work. [01:12:38] I still think we should split object equality so that we have one including a type checker, but if it's slow, it should hardly ever be used. So maybe it should be the one sent to a new ZID? [01:24:00] I’m hoping Z803 will be fixed by T363623 so we won’t need to reify an object just to identify its type! (re @Toby: I still think we should split object equality so that we have one including a type checker, but if it's slow, it should hardly e...) [12:10:44] How should python say what kind of items are in its list? In this implementation there's some problem with the serialising: Z16357: Actual result: { "Z1K1": "Z5", "Z5K1": "Z500", "Z5K2": { "Z1K1": { "Z1K1": "Z7", "Z7K1": "Z885", "Z885K1": "Z500" }, "Z500K1": "Could not serialize input Python object: 11" } } [P.S. I know I'm not meant to be using it yet, so just tell me [12:10:44] to wait if you like.] [12:30:43] 🤔 I think that’s T359233. You can see my hackaround in Z12863. (re @Toby: How should python say what kind of items are in its list? In this implementation there's some problem with the serialising: Z163...) [16:16:32] …actually, you can ignore the hackaround. It looks like this is conversion from the resulting integers into months. (re @Toby: How should python say what kind of items are in its list? In this implementation there's some problem with the serialising: Z163...) [20:40:45] Newsletter #156: New type: Gregorian calendar month -- https://www.wikifunctions.org/wiki/Wikifunctions:Status_updates/2024-05-22 [20:41:05] And I've seen that some of you have already started using it! Thank you! Please let us know if you encounter any issues [20:46:07] “And although both of them look very different – they use the identity operator of the given language to compare the two values – they work quite a bit differently” – is that meant to say that they look very similar? 😇 [20:50:14] It is! Thank you, fixed on wiki (sorry Email readers) [20:50:19] yay ^^ [20:52:54] And @amire80 , here's a thing that might help with translating: if you use language code on qqx now, it should consistently show either the ZID of object being displayed, or the key for the message. That should make it much easier to translate things! [20:53:36] and there are actually four implementations of same month 😉 [20:53:57] Do you have an example? (re @vrandecic: And @amire80 , here's a thing that might help with translating: if you use language code on qqx now, it should consistently show...) [20:55:08] https://www.wikifunctions.org/view/qqx/Z10000 (re @amire80: Do you have an example?) [20:55:36] ah, that's because I wrote it yesterday 🙂 Folks are two fast! 😄 (re @Al: and there are actually four implementations of same month 😉) [20:58:16] Z16359 is nice because it only uses built-in functions (or compositions of built-ins) (re @vrandecic: ah, that's because I wrote it yesterday 🙂 Folks are two fast! 😄) [20:59:10] https://tools-static.wmflabs.org/bridgebot/c7c4c3c1/file_61044_tgs.webp [20:59:29] ( *that* was a little bit bigger than I expected it to be, sorry )