Reduce memory thrashing when optimising large PDFs

[c2h5oh]

Before this change optimize.(*CombineDuplicateDirectObjects).Optimize has spent 97% of it's runtime in github.com/unidoc/unidoc/pdf/core.(*PdfObjectDictionary).WriteString and optimising a 500 page PDF with 4 images / page took a couple minutes on 8-thread cpu

After the change it's <1% and optimising the same file takes seconds.

[CLAassistant]

All committers have signed the CLA.

[codecov]

Codecov Report

Merging #347 into v3 will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##               v3     #347      +/-   ##
==========================================
+ Coverage   52.61%   52.61%   +<.01%     
==========================================
  Files         145      145              
  Lines       25235    25236       +1     
==========================================
+ Hits        13277    13278       +1     
  Misses      10170    10170              
  Partials     1788     1788

Impacted Files	Coverage Δ
pdf/core/primitives.go	75.15% <100%> (+0.05%)	⬆️

---

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 8d1cf1a...f03fb61. Read the comment docs.

[gunnsth]

Great, thanks.