Re: [PATCH] Teach pg_waldump to extract FPIs from the WAL - Mailing list pgsql-hackers

> > 6.
> > +        if (dir_status == 0 && mkdir(config.save_fpw_path, 0700) < 0)
> > Use pg_dir_create_mode instead of hard-coded 0007?
>
> I think I thought of that when I first looked at the patch ...  but, I'm
> not sure, since it says:
>
> src/include/common/file_perm.h-/* Modes for creating directories and files IN THE DATA DIRECTORY */
> src/include/common/file_perm.h:extern PGDLLIMPORT int pg_dir_create_mode;

Looks like it's pretty evenly split in src/bin:

$ git grep -o -E -w '(pg_mkdir_p|mkdir)' '**.c' | sort | uniq -c
      3 initdb/initdb.c:mkdir
      3 initdb/initdb.c:pg_mkdir_p
      1 pg_basebackup/bbstreamer_file.c:mkdir
      2 pg_basebackup/pg_basebackup.c:pg_mkdir_p
      1 pg_dump/pg_backup_directory.c:mkdir
      1 pg_rewind/file_ops.c:mkdir
      4 pg_upgrade/pg_upgrade.c:mkdir

So if that is the preferred approach I'll go ahead and use it.

> I was wondering if there's any reason to do "CREATE DATABASE".  The vast
> majority of TAP tests don't.
>
> $ git grep -ho 'safe_psql[^ ]*' '*pl' |sort |uniq -c |sort -nr |head
>    1435 safe_psql('postgres',
>     335 safe_psql(
>      23 safe_psql($connect_db,

If there was a reason, I don't recall offhand; I will test removing it
and if things still work will consider it good enough.

David