webhtml: reenable sort option for flamegraphs

[bbrks]

This sorts flamegraph elements lexicographically such that two similar profiles can be compared side-by-side to eyeball differences in flamegraph profile shapes and sizes.

Flamegraph elements prior to #367 were sorted, but since then, it has been difficult to eyeball 2 similar profiles side-by-side, as the elements appear in a somewhat random order.

Before sorting profile comparison:

After sorting profile comparison:

[nolanmar511]

I know there are differing opinions for whether the flame graph should be sorted by total value for each node, or whether it should be sorted by function name (mainly depending on the use case), but it looks like the order of nodes now is random. This change strikes me as positive (especially given that profile comparison is not supported in pprof's flame graph view).

@aalexand or @kalyanac -- This change LGTM, but I'll let either of you decide if this should be approved and merged,

[aalexand]

@nolanmar511 Yes, I think the change is an improvement. I am not sure whether this implementation of the flame graph is capable of sorting by the total value, but whatever sorting is better than no sorting.

Something is up with the CI tests though: looks like Travis CI is going to fail with

+++staticcheck github.com/google/pprof github.com/google/pprof/driver github.com/google/pprof/fuzz github.com/google/pprof/internal/binutils github.com/google/pprof/internal/driver github.com/google/pprof/internal/elfexec github.com/google/pprof/internal/graph github.com/google/pprof/internal/measurement github.com/google/pprof/internal/plugin github.com/google/pprof/internal/proftest github.com/google/pprof/internal/report github.com/google/pprof/internal/symbolizer github.com/google/pprof/internal/symbolz github.com/google/pprof/internal/transport github.com/google/pprof/profile github.com/google/pprof/third_party/d3 github.com/google/pprof/third_party/d3flamegraph github.com/google/pprof/third_party/svgpan
internal/driver/interactive_test.go:308:42: possible nil pointer dereference (SA5011)
The command "./test.sh" exited with 1.

which we need to fix and looks like AppVeyor build didn't start at all.

[codecov-io]

Codecov Report

Merging #491 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff            @@
##           master    #491      +/-   ##
=========================================
+ Coverage    67.1%   67.1%   +<.01%     
=========================================
  Files          37      37              
  Lines        7602    7603       +1     
=========================================
+ Hits         5101    5102       +1     
  Misses       2097    2097              
  Partials      404     404

Impacted Files	Coverage Δ
internal/driver/webhtml.go	100% <100%> (ø)	⬆️

---

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 5260658...75f3d0f. Read the comment docs.