Thread: Make error messages about WAL segment size more consistent
This started out as a small patch to make pg_controldata use the logging API instead of printf statements, and then it became a larger patch to adjust error and warning messages about invalid WAL segment sizes (IsValidWalSegSize()) across the board. I went through and made the primary messages more compact and made the detail messages uniform. In initdb.c and pg_resetwal.c, I use the newish option_parse_int() to simplify some of the option parsing. For the backend GUC wal_segment_size, I added a GUC check hook to do the verification instead of coding it in bootstrap.c. This might be overkill, but that way the check is in the right place and it becomes more self-documenting.
Attachment
Hi Peter, > This started out as a small patch to make pg_controldata use the logging > API instead of printf statements, and then it became a larger patch to > adjust error and warning messages about invalid WAL segment sizes > (IsValidWalSegSize()) across the board. Thanks for working on this. > I went through and made the > primary messages more compact and made the detail messages uniform. In > initdb.c and pg_resetwal.c, I use the newish option_parse_int() to > simplify some of the option parsing. For the backend GUC > wal_segment_size, I added a GUC check hook to do the verification > instead of coding it in bootstrap.c. This might be overkill, but that > way the check is in the right place and it becomes more self-documenting. I reviewed the code and tested it on Linux and MacOS with Autotools and Meson. The patch LGTM. -- Best regards, Aleksander Alekseev
On 22.08.23 16:26, Aleksander Alekseev wrote: > Hi Peter, > >> This started out as a small patch to make pg_controldata use the logging >> API instead of printf statements, and then it became a larger patch to >> adjust error and warning messages about invalid WAL segment sizes >> (IsValidWalSegSize()) across the board. > > Thanks for working on this. > >> I went through and made the >> primary messages more compact and made the detail messages uniform. In >> initdb.c and pg_resetwal.c, I use the newish option_parse_int() to >> simplify some of the option parsing. For the backend GUC >> wal_segment_size, I added a GUC check hook to do the verification >> instead of coding it in bootstrap.c. This might be overkill, but that >> way the check is in the right place and it becomes more self-documenting. > > I reviewed the code and tested it on Linux and MacOS with Autotools > and Meson. The patch LGTM. Thanks, committed.