[00:21:01] I think we can invite the user without them having to go through the captcha. @Sannita ? [02:00:37] I made a few for you. Z16855 is just a lookup. Z16862 should work, but barely fits within the time limits. Z16878 should take about double that time, so doesn't finish. These need to be sped-up before we can tackle anything beyond that, because every 10^3 would require an extra call to Z16878 . (re @Nicolas: we do have functions to turn number into letters in may [02:00:37] languages [02:00:38] but did anyone tried creating the reverse function? ie. letter...) [05:02:59] Has the wikilinks bot stopped working? [05:03:23] Z16855 Z16862 Z16878 [06:31:57] I'll check in around half an hour (re @Toby: Has the wikilinks bot stopped working?) [06:34:11] It works for other links, like Q5 and [[these links]], but not Wikifunctions links. Did anything changes with objects' json that I'm not aware of? [09:45:55] I've just created Z16904 for Bangla. Works for numbers < 10^9. (re @Nicolas: we do have functions to turn number into letters in may languages [09:45:56] but did anyone tried creating the reverse function? ie. letter...) [09:59:39] Very tidy! 👍 (re @myselftahmid: I've just created Z16904 for Bangla. Works for numbers < 10^9.) [11:03:21] @jhsoby and here it works... Did you fix it? (re @wikilinksbot: Z16904 – Bengali text to cardinal) [11:03:46] I shared with them the invite link again, cannot add them directly (re @vrandecic: I think we can invite the user without them having to go through the captcha. @Sannita ?) [11:04:19] no, nothing. Strange. Still doesn't work with the other ones (Z16878 et al) (re @Nicolas: @jhsoby and here it works... Did you fix it?) [11:18:38] Z16855 Z16862 Z16878 [11:18:40] The `<` signs were the problem, fixed now [17:38:22] That is a problem. We need a general rule. The Modern Standard Arabic rule was excellent. (re @wikilinksbot: Z16855 – English number <20 to natural number [17:38:23] Z16862 – English number <100 to natural number [17:38:25] Z16878 – English number <1000 to na...) [17:39:07] Although it was for a number to transcription conversion. [20:30:49] Check Z16919, it works in the general case. I didn't test it thoroughly though [20:33:55] It looks good. I think our edits clashed with no warning, however. Is that a known defect? (re @Ideophagous: Check Z16919, it works in the general case. I didn't test it thoroughly though) [20:36:16] no, I was testing it locally, then I copied my code to the implementation (re @Al: It looks good. I think our edits clashed with no warning, however. Is that a known defect?) [21:00:13] Z16926 facilitates testing: just edit it and put the same number in the call and the result validation (because we’re still waiting for T261460 😉) (re @Ideophagous: no, I was testing it locally, then I copied my code to the implementation) [23:28:05] Psst psst... There's a new type to try out: Z16927 - let me know if it works as expected. Following the proposal by [[User:Dolphyb]].