Created attachment 81435 [details] Sample document (as .odt, .html and .pdf) to show the issue. Also the three *.py files resulting from a cut&paste from each I can put Python program source code into a LibreOffice document and mark it as pre-Formatted (or Source Code). If I cut&paste this into a text editor from LibreOffice itself, it works. If I export it as an html file, and cut&paste from that, it works. But if I export it as a PDF file then a cut&paste of that will fail, as the PDF no longer retains the information about spaces - rather essential to pythign code. The problem also applies to other languages as in general any layout is lost, and spaces can be lost from inside string etc. I would like PDF exports to contain real spaces formatting in those document sections which are pre-formatted in just the same ay the the HTML export lists them as "<pre>" As a start, does anyone know which parts of the source actually handle these outputs for html and PDF? I could then try hacking around in the code myself, but at the moment I'm lost as to where this is (not) going on.
Created attachment 137143 [details] Example with wide character spacing I have run into what looks like a very closely-related issue in LibreOffice 5.2.2 - it seems to affect wide character spacing, not just pre-formatted text. In this example, spaces are *added* rather than removed. The HTML version does not have spaces added. Interestingly, spaces are not inserted between all characters: A R EA S O F E XP E RT I SE This breaks copy/paste, pdf2text and searching on the generated PDF.