[PATCH] perf: add another command to test performance of writing to output streams

[Yuya Nishihara]

On Mon, 22 Jun 2020 13:13:04 +0200, Manuel Jacob wrote:
> On 2020-06-22 12:58, Yuya Nishihara wrote:
> > On Mon, 22 Jun 2020 02:02:34 +0200, Manuel Jacob wrote:
> >> # HG changeset patch
> >> # User Manuel Jacob <me at manueljacob.de>
> >> # Date 1591314853 -7200
> >> #      Fri Jun 05 01:54:13 2020 +0200
> >> # Node ID 1d9781c4f9bcd2df5745b7e9d880dbd5b46c559e
> >> # Parent  61719b9658b156548634c216ffb2b9188266293b
> >> # EXP-Topic perfwrite2
> >> perf: add another command to test performance of writing to output 
> >> streams
> >> 
> >> The command was used recently while finding a solution for a buffering 
> >> bug
> >> (eventually fixed in f9734b2d59cc (the changeset description uses a 
> >> different
> >> benchmark)).
> >> 
> >> Calling the new command "perfwrite2" to keep the existing command 
> >> "perfwrite"
> >> unchanged is not very clever, but I didn’t have a better idea.
> > 
> > Maybe it's okay to remove the current perfwrite command so long as it 
> > can
> > be achieved by using the new implementation.
> 
> It would be `perfwrite2 --nlines=100000 --nitems=1 --item='Testing write 
> performance' --batch-line`.
> 
> Do you think it would be a good idea to replace the old command (using 
> the same name), but doing something different by default?

Yes. I don't think perfwrite is used as a source parameter of performance
testing, and even if it were, the command argument can be easily adjusted.

The change can be documented in relnotes.