Hi Slavi, Thanks a lot for the detailed review of this patch-set!I am about to send version 2 of NumPy interface patches which tries to address all the problems that you found.
Please review the new version carefully and be as critical as possible. Thanks! Yordan On 28.03.19 г. 14:47 ч., Slavomir Kaslev wrote:
On Wed, Mar 27, 2019 at 6:04 PM Yordan Karadzhov <ykaradzhov@xxxxxxxxxx> wrote: [...]diff --git a/kernel-shark/bin/sched_wakeup.py b/kernel-shark/bin/sched_wakeup.py[...]+ks.open_file(fname) +ofst, cpu, evt, pid, ts = ks.load_data()Returning a tuple of 5 numpy arrays as API will be impossible to extend in the future without breaking existing code. It's also easy for users to get the order of the return values wrong. Have you thought of returning a container object that's holding the numpy arrays? Pandas DataFrame[1] comes to mind although I don't know if it's worth to take Pandas as dependency for this. [1] https://pandas.pydata.org/pandas-docs/version/0.23.4/generated/pandas.DataFrame.html -- Slavi
|