Re: [HACKERS] [COMMITTERS] pgsql: Preventive maintenance in advance of pgindent run.

Peter Eisentraut <peter.eisentraut@2ndquadrant.com> writes:
> On 5/17/17 11:37, Tom Lane wrote:
>> Concretely, how about the attached?  I don't think this looks any
>> worse than the current layout.

> The previous setup has been in place for years and has never been a
> problem.  The alternatives are all quite a bit worse.

No, the previous setup hasn't been "in place for years".  These programs
were only NLS-ified last fall.  Before that the code looked like, eg,
printf("  -z, --stats[=record]   show statistics instead of records\n");printf("                         (optionally,
showper-record statistics)\n"); 

so that there weren't string continuations for pgindent to fool with.

I'm not really convinced that having usage() print two- or three-line
switch descriptions is "quite a bit better" than what I suggested.
        regards, tom lane