On Fri, 17 Feb 2023 at 21:44, David Rowley wrote:
> Updated patch attached.
After making another couple of small adjustments, I've pushed this.
Thanks for the review.
David
Thank you for having a look at this.
On Fri, 17 Feb 2023 at 14:01, Andres Freund wrote:
> > +set_ps_display_suffix(const char *suffix)
> > +{
> > + size_t len;
>
> Think this will give you an unused-variable warning in the PS_USE_NONE case.
Fixed
> > +#ifndef PS_USE_NONE
> > + /*
Hi,
On 2023-02-16 14:19:24 +1300, David Rowley wrote:
> After fixing up the set_ps_display()s to use set_ps_display_with_len()
> where possible, I discovered some not so nice code which appends "
> waiting" onto the process title. Basically, there's a bunch of code
> that looks like this:
>
>
While doing some benchmarking of some fast-to-execute queries, I see
that set_ps_display() popping up on the profiles. Looking a little
deeper, there are some inefficiencies in there that we could fix.
For example, the following is pretty poor:
strlcpy(ps_buffer + ps_buffer_fixed_size,
