On Sat, Dec 13, 2025 at 12:07:35PM +0700, Andrey Rudometov wrote:
> A patch with my approach to replacement is in attachments
> (to be applied after yours).
>
> p.s. I wonder why this builds at all - as far as I have tried, using
> BE in FE usually leads to a ton of compilation errors or, at least,
> warnings treated as errors - p.e, in this case backend definition
> uses ereport(), which does not work in frontend utilities.
Yep, that's indeed something that the patch should not do at all.
> - durable_rename(filename, filename_with_original_contents, FATAL);
> + err = durable_rename(filename, filename_with_original_contents);
> + if (err)
> + pg_fatal("could not rename file \"%s\"", filename);
I don't see a point in re-emitting an error message as well as you are
suggesting here. durable_rename() does this job already on failure
with a set of pg_log_error(). The only difference is that pg_fatal()
is a shortcut for pg_log_error()+exit().
--
Michael
