Fabien COELHO <coe...@cri.ensmp.fr> writes: > Attached v4 simplifies the format and fixes this one.
I think this goes way way overboard in terms of invasiveness. There's no need to identify individual call sites of PSQLexec. We didn't have anything like that level of detail before, and there has been no field demand for it either. What I had in mind was basically to identify the call sites of echoQuery, ie distinguish user commands from psql-generated commands with labels like "QUERY:" vs "INTERNAL QUERY:". We don't need to change the APIs of existing functions, I don't think. It also looks like a mess from the translatibility standpoint. You can't expect "%s QUERY" to be a useful thing for translators. regards, tom lane