Re: Should we automatically run duplicate_oids?

Peter Eisentraut <peter_e@gmx.net> writes:
> On 7/9/13 10:40 AM, Andrew Dunstan wrote:
>> This is actually a pretty trivial task. Here is a simple perl version:

> But that would make Perl a hard build requirement.

My opinion is that this script should only run if you've changed some
catalog .h files.  Requiring Perl in those cases is no worse than what
we already require it for (cf the keyword crosscheck script mentioned
upthread).
        regards, tom lane