On Wed, Jan 7, 2026 at 9:03 AM Tomas Glozar <[email protected]> wrote:
>
> út 6. 1. 2026 v 14:42 odesílatel Wander Lairson Costa
> <[email protected]> napsal:
> >
> > A few functions duplicate the logic for handling threshold actions.
> > When a threshold is reached, these functions stop the trace, perform
> > actions, and restart the trace if configured to continue.
> >
> > Create a new helper function, common_restart(), to centralize this
> > shared logic and avoid code duplication. This function now handles the
> > threshold actions and restarts the necessary trace instances.
> >
> > Refactor the affected functions main loops to call the new helper.
> > This makes the code cleaner and more maintainable.
> >
>
> The deduplication idea is good, but I find the name of the helper
> quite confusing. The main function of the helper is not to restart
> tracing, it is to handle a latency threshold overflow - restarting
> tracing is only one of possible effects, and one that is only applied
> when using --on-threshold continue which is not the most common use
> case. Could something like common_handle_stop_tracing() perhaps be
> better?
>
Sure, I will change the name in v3.
> > +enum restart_result {
> > + RESTART_OK,
> > + RESTART_STOP,
> > + RESTART_ERROR = -1,
> > +};
>
> Do we really need a separate return value enum just for this one helper?
>
If it was success/failure type of return value, we wouldn't need.
However, a three state code, I think it is worth for code readiness.
Do you have something else in mind?
> Tomas
>
