ARROW-5117: [Go] fix panic when nil or empty slices are appended to builders

[sbinet]

No description provided.

[sbinet]

PTAL

[sbinet]

@stuartcarnie: could I pick your brain on this?
it seems to me this is a valid change, but I'd like to know whether I didn't overlook anything...
(all tests pass, but...)

[codecov-io]

Codecov Report

Merging #4131 into master will decrease coverage by 12.25%.
The diff coverage is 97.05%.

@@             Coverage Diff             @@
##           master    #4131       +/-   ##
===========================================
- Coverage    87.8%   75.54%   -12.26%     
===========================================
  Files         745       61      -684     
  Lines       91468     4229    -87239     
  Branches     1252        0     -1252     
===========================================
- Hits        80309     3195    -77114     
+ Misses      11042      926    -10116     
+ Partials      117      108        -9

Impacted Files	Coverage Δ
go/arrow/array/fixedsize_binarybuilder.go	91.02% <100%> (+0.23%)	⬆️
go/arrow/array/numericbuilder.gen.go	92.87% <100%> (+24.7%)	⬆️
go/arrow/array/booleanbuilder.go	73.61% <100%> (+0.75%)	⬆️
go/arrow/array/binarybuilder.go	89.65% <50%> (-1.92%)	⬇️
go/arrow/memory/memory_avx2_amd64.go	0% <0%> (-100%)	⬇️
go/arrow/math/uint64_avx2_amd64.go	0% <0%> (-100%)	⬇️
go/arrow/math/int64_avx2_amd64.go	0% <0%> (-100%)	⬇️
go/arrow/math/float64_avx2_amd64.go	0% <0%> (-100%)	⬇️
go/arrow/memory/memory_amd64.go	28.57% <0%> (-14.29%)	⬇️
go/arrow/math/math_amd64.go	31.57% <0%> (-5.27%)	⬇️
... and 693 more

---

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 1ff1bca...d5b04bf. Read the comment docs.

[sbinet]

PTAL.

[sbinet]

gentle ping

[asubiotto]

LGTM minus my comment about the testing, although you might want to get someone else's input on this since I am new to the code.

[sbinet]

ping @alexandreyc @stuartcarnie

[alexandreyc]

LGTM modulo my comment on the test name for string builder.

[sbinet]

ok, thanks.
just waiting a couple of days for @stuartcarnie, to make sure I haven't overlooked anything in the logic of the "append" of arrays values (especially the exit-early path...) that could have adverse effects elsewhere.