On 11.08.25 08:04, Peter Eisentraut wrote:
> So here is an initial POC patch. I have written a script to convert a
> new src/backend/utils/misc/guc_parameters.dat to what would be
> guc_tables.c, but in the patch it's guc_tables_new.c. The
> guc_parameters.dat in the patch is populated only with a few entries
> that cover most of the different types and variants and possible
> settings, so we can see what it would look like. Eventually, this would
> require a big conversion.
>
> My next goal would be to make this work so that most of guc_tables.c is
> generated, and nothing else changes beyond that.
Ok, I did the big conversion, and tidied everything up so that it now
generates the big tables in guc_tables.c from the .dat file. This
basically works now.
Some notes:
- I included a preparatory patch to clean up some formatting in
guc_tables.c so that it's easier to visually compare the generated code.
- Similarly, for this time, I left the order of the entries in
guc_parameters.dat to be the same as it was in guc_tables.c, so that
it's easier to compare. Eventually I would like to re-order those,
probably alphabetically, but that can be a separate follow-up patch.
- In this patch, I left the code to be replaced in guc_tables.c, so that
it's easier to keep this patch rebased. But you should imagine that the
parts under NOT_USED will get deleted.
- One thing I didn't reproduce in the generated code is the line breaks
in the description strings. But that should be ok either way.
- In a few places there were complicated #ifdef's to determine the
default value. I extracted that out into separate symbols.
- Eventually it would be interesting to try to generate the other parts
of guc_tables.c, like the enums and the categories, but I'm leaving that
for another time.
- I moved the C code comments into the .dat file as Perl comments. I
figure that makes sense as that's where people would be editing now.
I feel that if people like this format, this is functional and useful as
a first step.
