Re: [COMMITTERS] pgsql: test_pg_dump TAP test whitespace cleanup

* Alvaro Herrera (alvherre(at)2ndquadrant(dot)com) wrote:
> Stephen Frost wrote:
> > * Alvaro Herrera (alvherre(at)2ndquadrant(dot)com) wrote:
>
> > > This will be undone by the next perltidy run.
> >
> > Ugh.
> >
> > I certainly hope what was there before wasn't the result of a perltidy
> > run as it was quite ugly and inconsistent..
>
> Maybe it was. I checked the diff after running perltidy after your
> commit and there are some changes that look like a reversion of your
> changes, but I don't know if there are other changes.

Yeah, I took a look at what perltidy is doing and at least some of that
was stuff I went through and fixed.

The changes to the %tests structure are pretty reasonable and it looks
like it's going to preserve at least some of what I did there because it
doesn't (always, at least) re-combine lines that have been split.
Unfortunately, what's it's doing to the %pgdump_runs structure are
rather annoying because it wants to turn something like:

- column_inserts => {
- dump_cmd => [
- 'pg_dump',
- '-f', "$tempdir/column_inserts.sql",
- '-a',
- '--column-inserts',
- 'postgres',
- ], },

into:

+ column_inserts => {
+ dump_cmd => [
+ 'pg_dump', '-f',
+ "$tempdir/column_inserts.sql", '-a',
+ '--column-inserts', 'postgres',
+ ],
+ },

I don't mind the change in indentation, of course, but obviously it's a
lot nicer when the '-f $tempdir/column_insert.sql' is all on the same
line since that's clearly a flag with an argument.

I could change all of those to be long versions, eg "--file=blah", to
keep them together. Not really ideal, but I'm not really worried that
getopt() is going to break either.

Unless anyone has a better idea, I'll see about doing that, and then
run perltidy of it to clean it up.

Thanks!

Stephen