[Cocci] Increase precision for display of SmPL script execution durations

[SF Markus Elfring]

Hello,

The Coccinelle software can display profiling results for the execution
of SmPL scripts. The measured durations are formatted in a way so that
the information “0.000 sec” is presented for some functions so far.
I guess that these data processing efforts were performed with a non-zero
time value.
How do you think about to show this detail precisely?

Regards,
Markus

[Julia Lawall]

Post by SF Markus Elfring
Hello,
The Coccinelle software can display profiling results for the execution
of SmPL scripts. The measured durations are formatted in a way so that
the information “0.000 sec” is presented for some functions so far.
I guess that these data processing efforts were performed with a non-zero
time value.
How do you think about to show this detail precisely?

Normally, one uses --profile to find the things that take seconds, not
to find the thing that take less than 0.01 seconds.

julia

Post by SF Markus Elfring
Regards,
Markus
_______________________________________________
Cocci mailing list
https://systeme.lip6.fr/mailman/listinfo/cocci

[Julia Lawall]

Post by Julia Lawall

Post by SF Markus Elfring
How do you think about to show this detail precisely?

Normally, one uses --profile to find the things that take seconds,
not to find the thing that take less than 0.01 seconds.

These values might look negligible.
I find the display of the information “0.000 sec” inappropriate.
https://github.com/coccinelle/coccinelle/blob/445c214e755dc74fae1d205d686365767ee64e85/commons/common.ml#L648
An adjustment (of a format string) could be so simple like the following.
diff --git a/commons/common.ml b/commons/common.ml
index 7cbeceaf..ca5399ad 100644
--- a/commons/common.ml
+++ b/commons/common.ml
@@ -657,7 +657,7 @@ let profile_diagnostic () =
pr "profiling result";
pr "---------------------";
xs +> List.iter (fun (k, (t,n)) ->
- pr (Printf.sprintf "%-40s : %10.3f sec %10d count" k !t !n)
+ pr (Printf.sprintf "%-40s : %f sec %10d count" k !t !n)

I know that the change is simple. But I don't want to see so many digits,
most of which are meaningless. In 46.742592, I'm only interested in the
46.

julia

)
)
Regards,
Markus