Re: Lazy printing in tabulated-list-mode

[Artur Malabarba <bruce.connor.am@gmail.com>]

Firstly, I don't really plan on implementing this lazy-printing
anymore, I'm just trying to optimize the printing a bit now.

Secondly, how do you see the running time on the profiler? I don't
know why it would report 1.3s, given that if I just evaluate
(benchmark-run 10 (tabulated-list-print)) I get a result between 6 to
10 seconds total (depending on whether package-hiding is on or off),
which means at least 0.6 seconds per run (of course).
Maybe the report I sent you was accidentally gathered on a small
package-menu (with only Gelpa configured).

0.6s per print is less than the "couple of seconds" I reported
earlier, because that was without byte-compiling. Still, I think it's
something we can optimize because the benchmark reports 67% of the
time was spent garbage collecting.



2015-05-04 1:19 GMT+01:00 Stefan Monnier <monnier@iro.umontreal.ca>:
>> Yep, that's what I was doing (if you consider 7 ~ 10).
>> Here's the command I was running under M-:
>
>> (progn (profiler-start 'cpu)
>> (tabulated-list-print)(tabulated-list-print)(tabulated-list-print)(tabulated-list-print)(tabulated-list-print)(tabulated-list-print)(tabulated-list-print)
>> (profiler-report) (profiler-stop))
>
> Hmm... but the numbers I saw seemed to indicate a running time of 1.3s,
> so that would mean you saw a running time of of 0.2s per call to
> tabulated-list-print which seems not to justify the addition
> of laziness.
>
>
>         Stefan