Closed julienw closed 5 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 88.45%. Comparing base (
50798ee
) to head (3d39ffa
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
And yet another follow-up: Maybe we should remove the self time bars altogether. Half the bars are useless for call nodes that don't have any self time. And for call nodes that do have self time, in most cases, the "self" samples all have the same category + subcategory. This means that at most one "self" bar is non-zero.
Got another idea about the border: using a border-right instead of a border-top/bottom:
wdyt @mstange ?
Another follow-up: I think all meter bars in the tooltip should use the same scale. In the deploy preview, the length of the bar in a subcategory is relative to its category's value. This is very misleading.
Yeah I was also wondering about that (see the PR description). 2 options:
Because it's so easy to change I'd like to get this right in this PR. So here are a few screenshots to compare all these options.
Current deploy preview:
Option 1:
Option 2:
I think I'd rather go with option 1. Then there are many things to do in follow-ups (possibly):
I pushed option 1 to the PR, so that it's ready to be merged if we all agree on that.
The changes
As we discussed several times, with this PR we now use the categories information in the flame graph tooltip.
This also changes some logic and CSS.
Here are some more thoughts about point 2:
Some screenshots
Before:
Without the CSS changes:
After, with all CSS changes:
Here is also an example with some self time:
Deploy preview
Deploy preview / production