[01:18:04] Thanks for noting this! I was able to fix Z26381, by specifying the language in the result validation (where it compares monolingual text objects). The language was simply missing. Obviously shouldn't cause such a drastic and unhelpful error; I will create a ticket for this. Z26380 is currently a duplicate of Z26381; haven't fixed it yet. (re @u99of9: I'm currently [01:18:04] getting fa [01:18:06] tal exceptions at Z26380 and Z26381. I don't know what they are, I was just patrolling recent changes. [...) [01:19:24] David At Z6884 why don't we specify that the list must be typed Wikidata item reference? : https://tools-static.wmflabs.org/bridgebot/c513c918/file_72057.jpg [01:29:36] Because we figured we might as well support all of the Wikidata reference types, and that was pretty easy to do. I think it likely will be useful sometimes to have a enum of lexeme references, or lexeme form references, or property references, for some purpose. (re @u99of9: David At Z6884 why don't we specify that the list must be typed Wikidata item reference?) [01:33:40] I've just pitched the initial proposal for [[Wikifunctions:Type proposals/Chemical element]] which I've been pondering for months. [06:01:19] David Al just a heads up that the undefined values are gradually propagating issues downstream. For example, I just had to disable most of the implementations of Z19892 due to Z26532 (which in turn was ultimately required by the validation of Z26529). [06:58:57] 🙄 Lucky I made so many implementations! I think this should be trapped further upstream, however. That is, the void should be allowed in the Wikidata object but handled in the default extraction functions. For example, Z25300 could return an upper bound equal to Z6010K1/amount. (re @u99of9: David Al just a heads up that the undefined values are gradually propagating [06:58:57] issues dow [06:58:57] nstream. For example, I just had to disab...) [07:09:39] I'll have a think about this, but does void always mean the value is exact, or could it mean the bound is unknown? (re @Al: 🙄 Lucky I made so many implementations! I think this should be trapped further upstream, however. That is, the void should be al...) [07:27:25] No, void only means that there is no value on Wikidata. In practice, though, our default interpretation is that the bound is equal to the amount. A different function could make a different assumption, returning an object other than Rational number (I would hope). (re @u99of9: I'll have a think about this, but does void always mean the value is exact, or could it mean the [07:27:25] bound is unknown?) [08:18:40] While working on implementations, I keep getting the error "Unable to run tests. Please reload" when trying to check the connected tests. This issue occurs not just in one function but in every implementation that I am editing, all tests fail with the same error. [08:47:53] It's a problem with JavaScript implementations (re @baibaswata: While working on implementations, I keep getting the error "Unable to run tests. Please reload" when trying to check the connect...) [12:26:18] 0237 [23:17:00] As far as I've seen it is correct though. Once you save and reload the page, the tests run correctly. (re @baibaswata: While working on implementations, I keep getting the error "Unable to run tests. Please reload" when trying to check the connect...)