Improve boundschecks during assemble

[KnutAM]

I think this should be safer, but I haven't benchmarked if there is a performance regression.

[codecov-commenter]

Codecov Report

Patch coverage: 100.00% and no project coverage change.

Comparison is base (c6c185c) 92.61% compared to head (2d0effb) 92.61%.

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #706   +/-   ##
=======================================
  Coverage   92.61%   92.61%           
=======================================
  Files          30       30           
  Lines        4430     4430           
=======================================
  Hits         4103     4103           
  Misses        327      327           

Impacted Files	Coverage Δ
src/assembler.jl	99.34% <100.00%> (ø)

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[fredrikekre]

I suppose requiring the same length is enough if we instead do

for (i, j) in zip(keys(ge), keys(dofs))
    addindex!(g, ge[i], dofs[j])
end

but in practice you would never utilize this extra degree of freedom I don't think.