Robert Haas <robertmhaas@gmail.com> writes:
> On Wed, Apr 27, 2016 at 2:23 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Um, we normally take the buildfarm's list of typedefs, not anything
>> manually created.
> Well, we can still do that, but I don't see much advantage in it. It
> just churns the file to the extent that manual review of the changes
> is impossible, and then when pgindent does the wrong thing it only
> gets reported after the fact. How is that better than making sure
> that the contents of the file are such as to actually produce good
> output from pgindent?
On what grounds do you claim the buildfarm result is unstable?
I've been using that for a long time and it works fine. Moreover,
ignoring that data is a bad idea because it reflects platform-specific
variations in the set of typedefs that are known. If you build a
typedefs list based only on what works on your machine, it likely
won't work for other people.
regards, tom lane
