Network track for the timeline #1189
Conversation
|
There seems to be an odd first-paint effect where the whole canvas is filled with requests: |
|
Wheel scrolling doesn't seem to work while the mouse is over the network track. |
@mstange: I can't reproduce this. I tried recording my screen, and then loading both from the add-on, viewing existing profiles, and committing new ranges. |
Codecov Report
@@ Coverage Diff @@
## master #1189 +/- ##
==========================================
+ Coverage 75.9% 76.06% +0.15%
==========================================
Files 144 145 +1
Lines 9364 9501 +137
Branches 2319 2350 +31
==========================================
+ Hits 7108 7227 +119
- Misses 2013 2031 +18
Partials 243 243
Continue to review full report at Codecov.
|
| expect(getRightClickedTrack(getState())).not.toEqual(trackReference); | ||
| expect(getSelectedThreadIndex(getState())).not.toBe(threadIndex); | ||
| expect(getLocalTrackRow().hasClass('selected')).toBe(false); | ||
| describe('with a thread track', function() { |
gregtatum
Aug 10, 2018
Author
Member
I ended up taking Nicolas Chevobbe's review advice here and re-organized these test files, as I was feeling lost while working on the tests. The setup functions are located at the bottom of the page.
I would recommend turning on "hide whitespace changes" in the Diff Settings on this review, as there is an indentation change.
I ended up taking Nicolas Chevobbe's review advice here and re-organized these test files, as I was feeling lost while working on the tests. The setup functions are located at the bottom of the page.
I would recommend turning on "hide whitespace changes" in the Diff Settings on this review, as there is an indentation change.
| view.update(); | ||
| expect(view.find('.timelineTrackHidden').exists()).toBe(true); | ||
| expect(view.find('.timelineTrack').exists()).toBe(false); | ||
| }); | ||
| }); | ||
|
|
gregtatum
Aug 10, 2018
Author
Member
The following are the new tests:
The following are the new tests:
This is fixed in the latest code. |
|
This PR looks really good. :) Besides the one question above, I realized that the markers in the network panel get updated when selecting a range, but the list is not updated (see screenshot, checked with following profile)
When you select a range for the marker-table, the markers as well as the list get updated. ======== When selecting just the network thread, the shadow makes it a bit hard so see the network markers in the thread. It also feels dense.
======== How would you feel about adding one more row or some additional padding at the bottom of the network thread (so increase the min-height), as the thread seems to get "swallowed" when it does not contain markers?
|
| type: 'Network', | ||
| id, | ||
| pri: 0, | ||
| status: 'DONE', |
zoepage
Aug 17, 2018
Contributor
The status DONE got me confused as I thought we just have START and STOP, so I thought I rather check in if there is something I am missing? :)
The status DONE got me confused as I thought we just have START and STOP, so I thought I rather check in if there is something I am missing? :)
gregtatum
Aug 20, 2018
Author
Member
It was just mock data, I didn't remember what the strings were off the top of my head, since they aren't in our types. I'll change it to STOP to be consistent and not confusing.
It was just mock data, I didn't remember what the strings were off the top of my head, since they aren't in our types. I'll change it to STOP to be consistent and not confusing.
The shadow is created by the OverflowEdgeIndicator component. It should only show up when the timeline can scroll down, which this is not doing in your screenshots. I narrowed it down to an STR. I think we should fix this in a follow-up, I filed that as #1205.
This is #865.
Yeah, I can throw another row in there for that case. I feel like the network track being empty is a bit weird, although I think having it disappear might be confusing. Perhaps having it hide itself is a better option here? |
+1 in hiding itself when it is empty. Do you want to file a follow up for this? In case we have just very few markers and the row is still min-height, we can re-check again when it landed and file a follow up bug, if this still feels like it gets "swallowed". Thank you for the changes :) |
Deploy preview
I'm pushing this up at the end of the day, feel free to look at the deploy preview, but I plan on reading through it all again with some fresh eyes tomorrow before asking for review.Edit: This is now ready for review, @ola, would you feel comfortable reviewing this? I have split the work into multiple logical commits, so it will be easiest to look at each in turn, rather than the whole at once. I still haven't reproduced the first-paint effect Markus describes below.
Also resolves #1196