On 06/05/2018 04:57 PM, Andres Freund wrote: > But we don't really change the > collation of function invocations per-call.
Is that true? (Not a rhetorical question; I'm unsure.) Is your argument that the expression's collation is determined once for each call /site/, and thereafter doesn't change over repeated invocations via any one call site, and a unique FmgrInfo is populated at each call site? This is at the edges of my current understanding of how those gears turn, so it could very well be true, but it seems to me that's how the question would have to be put. -Chap