2012/4/22 Németh Márton <[email protected]>:
> Thomas Spura wrote:
>> 2012/4/21 Németh Márton <[email protected]>:
>>> diff --git a/main.cpp b/main.cpp
>>> index 0e57ee1..67db2eb 100644
>>> --- a/main.cpp
>>> +++ b/main.cpp
>>> @@ -240,6 +240,7 @@ void report(int time, int iterations, char *file)
>>> one_measurement(time);
>>> report_show_tunables();
>>> finish_report_output();
>>> + clear_tuning();
>>> }
>>> /* and wrap up */
>>> learn_parameters(50, 0);
>>> @@ -415,6 +416,7 @@ int main(int argc, char **argv)
>>> learn_parameters(500, 0);
>>> save_parameters("saved_parameters.powertop");
>>> end_pci_access();
>>> + clear_tuning();
>>> reset_display();
>>>
>>> clear_all_devices();
>>
>> These two chunks don't apply here. Did you create the patch against
>> currend HEAD commit 1dfdb80d? But look fine, when applying manually.
>
> Yes, I used 1dfdb80d revision as a base. I don't know what could have went
> wrong.
Could also be my fault, as I'm pretty new to the mailinglist workflow.
(Previously, I only did github pull requests etc).
>>> diff --git a/tuning/tuning.cpp b/tuning/tuning.cpp
>>> index a0c3ffa..6a359ae 100644
>>> --- a/tuning/tuning.cpp
>>> +++ b/tuning/tuning.cpp
>>> @@ -312,3 +312,15 @@ void report_show_tunables(void)
>>> fprintf(reportout.csv_report,"\n");
>>> }
>>> }
>>> +
>>> +void clear_tuning()
>>> +{
>>> + while (!all_tunables.empty()) {
>>> + delete all_tunables.back();
>>> + all_tunables.pop_back();
>>> + }
>>> + while (!all_untunables.empty()) {
>>> + delete all_untunables.back();
>>> + all_untunables.pop_back();
>>> + }
>>> +}
>>
>> I'd prefer a for loop like in commit 1dfdb80d. Easier to read and
>> should be faster than poping and always check for empty().
>
> Are you speaking about this commit?
Same commit id as above... Copy&paste error...
I was talking about commit 47cfb71 and e08a9c5, e.g.:
+void clear_all_devices(void)
+{
+ for (int i = 0; i < all_devices.size(); i++) {
+ delete all_devices[i];
+ }
+ all_devices.clear();
+}
Here anything is deleted in order and all_devices is resized once (and
not .size() times like in the tunables case).
Greetings,
Tom
_______________________________________________
Power mailing list
[email protected]
https://bughost.org/mailman/listinfo/power
