search

gojuno / composer

Reactive Android Instrumentation Test Runner. Archived. Marathon is recommended as an alternative (https://github.com/Malinskiy/marathon).
Apache License 2.0
546 stars 45 forks source link

Add search bar and filters for logcat output #147

Open tokou opened 6 years ago

tokou commented 6 years ago

There are many ways to approach this. I decided against adding to the window.test object as logcat output can be very large.

I kept the static html generated by the Jar. I just added an id static_logs in order to remove it once we load the logs in js.

I made a LogContainer component that shows the filters, the search bar and the logs once they are loaded from the file (it replaces the static logs).

I extracted the SearchBar component and made it generic and used by LogFilter and SuiteFilter

I tested this with a 10k lines logcat output file. Loading is seamless thanks to the html. Search is kinda laggy but acceptable. Maybe there's a React trick out there to make it faster.

yunikkk commented 6 years ago

Hey @navstyachka could you PTAL?)

navstyachka commented 6 years ago

Hi @tokou! Thank you for the contribution, but as soon as this is a massive change, I will need more time to review that.

tokou commented 6 years ago

I just noticed that it does not work when opening the index file directly in the browser. It works when you serve it from a local web server though. I will look into it.

tokou commented 6 years ago

I made an update to get the logs from the DOM tree instead of reading the logcat file. It works without a web server now.

tokou commented 6 years ago

The build seems to have failed for reasons unrelated to the code. Any way to retrigger it ?

yunikkk commented 6 years ago

@tokou I've triggered it

tokou commented 6 years ago

Hi @navstyachka Thanks for your detailed review. As you have guessed I am not really a react dev but an Android dev that wanted to contribute. I will take the time to make corrections based on your review. I understand that it is not possible to merge the PR even after the corrections. As I said in previous comments, I tried other approaches but they all resulted in performance problems when logs get large. I would be interested to explore other approaches if you have any suggestions.