Unexpected garbage collection when using append!

The issue’s re-appeared. This time it only manifests after importing the package Plots:

julia> using BenchmarkTools

julia> function motif2(v::Vector{Vector{Float64}})     
           out = Vector{Float64}(undef, sum(length, v))
           offset = 0
           for vect in v
               for i ∈ eachindex(vect)
                   @inbounds out[offset + i] = vect[i] 
               end
               offset += lastindex(vect)
           end
           out
       end
motif2 (generic function with 1 method)

julia> vects = [rand(10) for i = 1:10];

julia> @benchmark motif2($vects)
BenchmarkTools.Trial: 10000 samples with 983 evaluations.
 Range (min … max):   61.343 ns …  1.403 μs  ┊ GC (min … max): 0.00% … 65.96%
 Time  (median):      76.399 ns              ┊ GC (median):    0.00%
 Time  (mean ± σ):   106.685 ns ± 90.701 ns  ┊ GC (mean ± σ):  7.25% ±  9.22%

  ▃█▆▄▄▃▃▂▂▁▁▁▁▁                                               ▁
  ███████████████████▇▇▇▇▇▇▇▆▅▆▆▅▅▅▃▅▅▅▅▄▃▅▂▄▅▃▅▅▅▄▄▅▃▄▅▄▄▅▄▅▅ █
  61.3 ns       Histogram: log(frequency) by time       559 ns <

 Memory estimate: 896 bytes, allocs estimate: 1.

julia> using Plots

julia> @benchmark motif2($vects)
BenchmarkTools.Trial: 5878 samples with 978 evaluations.
 Range (min … max):   70.654 ns … 59.460 μs  ┊ GC (min … max):  0.00% … 99.52%
 Time  (median):      83.027 ns              ┊ GC (median):     0.00%
 Time  (mean ± σ):   871.062 ns ±  6.223 μs  ┊ GC (mean ± σ):  82.20% ± 11.33%

  █                                                            ▁
  █▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▄▇ █
  70.7 ns       Histogram: log(frequency) by time      54.2 μs <

 Memory estimate: 896 bytes, allocs estimate: 1.

However, unlike the previous issue, re-starting the REPL and not importing Plots fixes the issue.