On Fri, Apr 19, 2024 at 1:35 AM Michael Paquier <mich...@paquier.xyz> wrote: > Are you sure that relying on Temp::File is a good thing overall? The > current temporary file knowledge is encapsulated within Utils.pm, with > files removed or kept depending on PG_TEST_NOCLEAN. So it would be > just more consistent to rely on the existing facilities instead? > test_json_parser is the only code path in the whole tree that directly > uses File::Temp. The rest of the TAP tests relies on Utils.pm for > temp file paths.
Yeah, I think this patch invented a new solution to a problem that we've solved in a different way everywhere else. I think we should change it to match what we do in general. -- Robert Haas EDB: http://www.enterprisedb.com
