Re: proposal: possibility to read dumped table's name from file

Hi

> What are your thoughts on this version? It's not in a committable state
> as it
> needs a bit more comments here and there and a triplecheck that nothing was
> missed in changing this, but I prefer to get your thoughts before spending
> the
> extra time.
>

I think using pointer to exit function is an elegant solution. I checked
the code and I found only one issue. I fixed warning

[13:57:22.578] time make -s -j${BUILD_JOBS} world-bin
[13:58:20.858] filter.c: In function ‘pg_log_filter_error’:
[13:58:20.858] filter.c:161:2: error: function ‘pg_log_filter_error’ might
be a candidate for ‘gnu_printf’ format attribute
[-Werror=suggest-attribute=format]
[13:58:20.858] 161 | vsnprintf(buf, sizeof(buf), fmt, argp);
[13:58:20.858] | ^~~~~~~~~
[13:58:20.858] cc1: all warnings being treated as errors

and probably copy/paste bug

diff --git a/src/bin/pg_dump/pg_restore.c b/src/bin/pg_dump/pg_restore.c
index f647bde28d..ab2abedf5f 100644
--- a/src/bin/pg_dump/pg_restore.c
+++ b/src/bin/pg_dump/pg_restore.c
@@ -535,7 +535,7 @@ read_restore_filters(const char *filename,
RestoreOptions *opts)
case FILTER_OBJECT_TYPE_EXTENSION:
case FILTER_OBJECT_TYPE_FOREIGN_DATA:
pg_log_filter_error(&fstate, _("%s filter for \"%s\" is
not allowed."),
- "exclude",
+ "include",
filter_object_type_name(objtype));
exit_nicely(1);

Regards

Pavel

>
> --
> Daniel Gustafsson
>
>