[Week 01]— Getting Stuff Done!#
It’s time for Action ; Code-Sleep-Coffee Repeat!
This blog post marks my account for the first week of coding for GSoC’19. While it may not be as long as my previous posts, I would hope to make posts every week, so it would be as recent and relevant.
Making Progress for the Week#
While I had already made progress with the tasks, now it was time to get the unmerged PRs merged. I started to refine my PRs, and after going through lot of iterations, I decided to drop this idea and do something else. A fresh mind would be more receptive to find bugs.
Frankly speaking, I did not make any progress, even though in my mind I
thought I made 😛. I mostly had a lot of discussion with @Cadair
about the issue of failing tests of NDCube, and I spent a lot of
time understanding the plotting of NDCube. The plotting seems to be
very complex, and it contained small portions of code which made no
sense to me at all. All the time, I spent time understanding the code,
but made no progress.
Sometimes debugging is helpful in understanding the flow of code!
I had been spending time in understanding, more than making real progress. This is wrong as writing some code always help in getting unstuck. A lot of times people suggest to understand the code before making changes, but if you spend a lot of time understanding the flow of code without any progress, it is better to crack open your favourite editor and start using breakpoint to identify the flow of code and make notes (yes notes!) of the changes unexpected.
Another thing I did is use test-cases to identify how the input was getting modified. Sometimes, an example is really helpful to map the abstraction of a method into something practical. I used the tests to inspect the portions of code which were creating nuisance, as I submitted another PR about fixing most the issues. There are few test cases which are failing, so it would take more time before this PR is merged.
Link to previous post#
Contributing my way to SunPy1.0 — Link