[00:19:25] David can you have a think about whether this type would fit in with some of your plans: https://www.wikifunctions.org/wiki/Wikifunctions:Type_proposals/Wikifunctions_object_reference#Comments [01:30:05] Al it seems like some of the issues you listed in the proposal have already been solved in other ways? (re @u99of9: David can you have a think about whether this type would fit in with some of your plans: https://www.wikifunctions.org/wiki/Wiki...) [04:21:49] Something has gone wrong at Z25356. In the tests column the first fail has a "null" in the output. But then when it is run in Try this implementation, it works fine. I'll try some nudging now, but hope this is not a sign of a wider issue. : https://tools-static.wmflabs.org/bridgebot/1dba6a6f/file_74065.jpg [04:24:37] Phew. Nudging the composition itself fixed everything, but I'm still wary about why this went wrong in the first place. (re @u99of9: Something has gone wrong at Z25356. In the tests column the first fail has a "null" in the output. But then when it is run in Tr...) [05:56:56] This implies there's some asymmetry in how we handle display functions between the test output and the expectation. (re @u99of9: Phew. Nudging the composition itself fixed everything, but I'm still wary about why this went wrong in the first place.) [05:57:30] Is there a task tracking this? If not, can you make one with repro instructions (and your screenshot)? This is very weird. [06:09:48] It's quite common to see. The ones in Tests are coming from the cache. The one in "Try this" seems live. But it's hard to write long-lasting repro instructions because as soon as the cache is recalculated, the issue goes away. (re @wmtelegram_bot: Is there a task tracking this? If not, can you make one with repro instructions (and your screenshot)? This is [06:09:48] very weir...) [06:11:18] Oh, sorry, maybe you're talking about how 10+/-Null got into the cache in the first place. For that I have no idea. This is the first time I've seen it. (re @wmtelegram_bot: This implies there's some asymmetry in how we handle display functions between the test output and the expectation. (re ...) [06:35:05] I’ve been able to work around some of the issues by creating quote objects, for example, but those have their own problems and it’s all a bit flaky and sometimes pretty inefficient. I don’t think my proposal needs much development work; we should just do it! (re @u99of9: Al it seems like some of the issues you listed in the proposal have already been solved in other [06:35:05] ways?) [06:43:01] Great. Sorry I'm slow to the party, but I'm now dancing to your tune. (re @Al: I’ve been able to work around some of the issues by creating quote objects, for example, but those have their own problems and i...) [08:37:19] (I think ZReference is supposed to be available to code implementations, like ZPair now is, but it’s still “undefined”. I’ll file a ticket if Cory doesn’t correct me.) [22:39:33] Sure, I'll be happy to start thinking and bring it up with the team. The possibility came up before, but didn't seem like such a high priority at the time. The proposal makes some good points. (re @u99of9: David can you have a think about whether this type would fit in with some of your plans: https://www.wikifunctions.org/wiki/Wiki...) [22:46:39] yes. I kinda thought there was. Will make one now. (Also, apologies for not responding on this topic sooner.) (re @Al: There should probably be an open task to make it operational 🤷‍♂️)