[01:13:39] I've copied your formula at Z29362, which also seems to work. This is looking good. (re @Al: 🤔 All six tests pass for Z29349 but I’m leaving it disconnected until I have time to validate the impact, because it looks like ...) [01:38:27] Yes, also working for Z29365. This can go further, but should be enough to make the simplest tables directly without extra helper functions. Thanks Al (re @u99of9: I've copied your formula at Z29362, which also seems to work. This is looking good.) [02:22:01] Z29368 now works the way @dvd_ccc27919 assumed the other one would. (re @u99of9: Yes, also working for Z29365. This can go further, but should be enough to make the simplest tables directly without extra helpe...) [02:41:06] Which helps the new composition of Z29346 to be about 4 times faster than the hack. (re @u99of9: Z29368 now works the way @dvd_ccc27919 assumed the other one would.) [04:33:49] God bless the great minds in this blessed platform! [09:34:07] Thanks for running with this… (re @u99of9: Which helps the new composition of Z29346 to be about 4 times faster than the hack.) [10:54:44] Of course! Now that the roadblock is cleared, I have visions of a single auto_wiktionary_table(display_language, lexeme_id, element_function) (re @Al: Thanks for running with this…) [10:56:44] From the argument types of the element function we can get the enum_types whose elements define the rows and columns. (re @u99of9: Of course! Now that the roadblock is cleared, I have visions of a single auto_wiktionary_table(display_language, lexeme_id, elem...) [11:02:25] For that we will need things like Z29383 to work, but I can see that it is already close. (re @u99of9: From the argument types of the element function we can get the enum_types whose elements define the rows and columns.) [11:25:53] It depends what you mean by “work”… The function is expecting to return a full type rather than a reference, but the Z17K1 contains type references. (re @u99of9: For that we will need things like Z29383 to work, but I can see that it is already close.) [11:28:32] Either is probably enough. Do we have a different function to use in the validation of that test so that it passes? (re @Al: It depends what you mean by “work”… The function is expecting to return a full type rather than a reference, but the Z17K1 conta...) [11:50:03] One more tricky step would be to get from Z25340 to Q162378. I guess we can't do this at the moment? (re @u99of9: From the argument types of the element function we can get the enum_types whose elements define the rows and columns.) [11:51:55] Ah… I don’t think so. The returned type doesn’t look quite right… 🤔 (re @u99of9: Either is probably enough. Do we have a different function to use in the validation of that test so that it passes?) [11:57:13] No… in this case, the items’ P31s should give us this, but there’s nothing in the enum’s definition for the general case. (re @u99of9: One more tricky step would be to get from Z25340 to Q162378 (for the 0th column title). I guess we can't do this at the moment?) [12:48:48] It seems to be fine and it works in Z19084 🤷‍♂️ (re @Al: Ah… I don’t think so. The returned type doesn’t look quite right… 🤔) [12:50:59] Ah, that's not actually terrible, and I hadn't considered it. We have a function to get the items from the type. (re @Al: No… in this case, the items’ P31s should give us this, but there’s nothing in the enum’s definition for the general case.) [12:55:17] But it doesn't pass the (?too) strict tests... (re @Al: It seems to be fine and it works in Z19084 🤷‍♂️) [13:11:51] It passes in same Type and Z19108. I don’t know why they don’t work in the test case. (re @u99of9: But it doesn't pass the (?too) strict tests...) [13:18:53] (The other test fails because the code is trying to return a reference that is not a Type, although it is a valid reference to the correct type.) (re @Al: It passes in same Type and Z19108. I don’t know why they don’t work in the test case.)