On Fri, 2023-02-10 at 22:50 -0500, Robert Haas wrote: > The fact that you're figuring out how it all works from reading the > source code does not give me a warm feeling.
Right. On the other hand, the behavior is quite well documented, it was just the keyword that was undocumented (or I didn't find it). > > But it seems like a better place for us than libc for the reasons I > > mentioned in the other thread. > > It may be. But sometimes I feel that's not setting our sights very > high. :-( How much higher could we set our sights? What would the ideal collation provider look like? Those are good questions, but please let's take those questions to the thread about ICU as a default. The topic of this thread is: Given that we are already offering ICU support, should we canonicalize the locale string stored in the catalog? If so, should we use the ICU format locale IDs, or BCP 47 language tags? Do you have an opinion on that topic? If not, do you need additional information? -- Jeff Davis PostgreSQL Contributor Team - AWS