Fixes TraceTimelineViewer span details render issue

[rubenvp8510]

Which problem is this PR solving?

• This fixes Issue Span bar broken on click #628 , a regression when refactored VirtualizedTraceView to remove deprecated method componentWillReceiveProps and use the componentDidUpdate instead.

Short description of the changes

• The component was using componentWillReceiveProps to derive data from properties and store in the instance attributes to avoid recompute it in each render call, With the migration, that process was changed to componentDidUpdate, but this is not correct, RowState computation on that method doesn't trigger a new render process, so we ended up seeing something not in sync with the RowState. This change uses memoization for rowState to avoid recompute it each time.

[codecov]

Codecov Report

Merging #629 into master will increase coverage by 0.05%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #629      +/-   ##
==========================================
+ Coverage   93.54%   93.59%   +0.05%     
==========================================
  Files         227      227              
  Lines        5901     5902       +1     
  Branches     1485     1483       -2     
==========================================
+ Hits         5520     5524       +4     
+ Misses        340      337       -3     
  Partials       41       41              

Impacted Files	Coverage Δ
...ePage/TraceTimelineViewer/VirtualizedTraceView.tsx	95.08% <100.00%> (+2.51%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5b4ca26...ac10a5f. Read the comment docs.

[objectiser]

Tested with hotrod example, seems to fix the problem.

[yurishkuro]

@tklever please review

[tklever]

I pulled this code locally and I can no longer replicate the "empty box" problem. This looks good (to my limited expertise).

I cooked up a couple tests last night to lock in @rubenvp8510's work in this PR, specifically to ensure that the memoization he added is preserved and can't be accidentally removed. When this PR (or some version of it) lands, I'll submit some unit tests so we (or I in this case :-( ) can't reintroduce this regression in the future.

[rubenvp8510]

@yurishkuro Could you approve this change please? Thanks!