Jump to content

User:Bencmq/Progress

From mediawiki.org
(Redirected from User:Bencmq/Thanks)
7 Feb
8 Feb
  • Working implementation of flow-thanks

Issues

  • Getting post, workflow etc from post-id. i.e. what information to pass to flow-thank API call
  • getting creator id sometimes returns a string (for posts made before Friday) but now an integer
9 Feb

Issues

  • Styling on .mw-ui-disabled
11 Feb
  • Refactor css class handing on active and disabled link.
14 Feb
  • Fix issue with creator id sometimes being a string and sometimes an integer
  • Update backend to match new flow-thanks API format
16 Feb
  • FBOA video
  • Reading on browser testing on mw.org
19 Feb
  • Completed FBOA video
20 Feb
  • Reading on browser testing and API test (MediaWikiTestCase)
  • Playing with Selenium locally, starting on flow-thanks browser test.
24 Feb
  • Experimenting with Selenium, wrote small test for anon user
27 Feb
  • Writing APITest for Flow Thanks, currently stuck on preparing mock data
1 March
  • Some improvements for Flow patch based on code review
4 March
  • More study on page-object, wrote 2 more test cases

Tests can be found at

10 March
The browser test part can be found at: [1], looking for feedback
skipping scenarios conditionally: [2]
12 March

Integration test for API completed at [3]

15 March

Sorted out the CSS issue with mw-ui-button in core, turns out it's already fixed (now on gerrit:113066)

16 March

Apply relevant css classes to interaction links (now on gerrit:119812)

17 March

Start working on shifting flow's light button design to core

20 March

Side track: better timestamp for i18n [4]

26 March

https://gerrit.wikimedia.org/r/#/c/121101/

30 March
1 April
3 April
4 April
7 April
  • Looking into vagrant slowness
10 April
  • Get HHVM installed with help from wctaiwan's guide. Had minor improvements over the loading /dump
  • Also added some additional info to the guide that helped me getting it up and running.
    • More details:
      • Enabling cache does not seem to have an observable impact.
      • Switching to HHVM slows down first few loadings by around a second or so, but subsequent page loads are slightly faster. Some data in /dump, needs to present it clearer.
        • According to a friend working on HHVM, if on Linux (mediawiki-vagrant is), JIT will kick in after 7 times.
      • The main bottleneck is still loading of js and css files. The loading time of the whole page would be reduced by half if they are cached by browser but that's probably not something desirable in development.