OpenXML Filter: XLSX: improve memory consumption for identical cell data

[Denis Konovalyenko]

This is a follow-up for issue #984. Excel cell data are copied on extraction to allow contextualised and independent translations. So, the scope of this enhancement is to provide an alternative way to extract the data uniquely with possible loss of functionality for specifically marking rows and cells as metadata.

‌

[Denis Konovalyenko]

A related solution was provided in the scope of pull request #797.

[Denis Konovalyenko]

Pull request #797 was merged.

[Chase Tingley]

I’m not sure #984 is the right number for what this is a follow-up, to. To provide extra context, historically we have always extracted repeated cell content independently for each usage, rewriting the share strings table on the fly to allow for independent translations. However, for some use cases (notably, raw machine translation) this is overkill, so the goal was to add a way to disable it and just translate the data as it exists in the file. The default behavior is unchanged.