Header formatting is off when docx file is opened and converted in the writer v7.x

[omidghazara]

I am trying to resolve an issue regarding the imported docx files from word office 365 to the redhat linux server that uses LO for conversion of these docx files. The issue is that when opening the file, the header and some formatting regarding the fonts and other styles disappear in the newer versions of LO. For example the logo in the header is rendered but now shown and is malpositioned in the middle of the header and some texts also. These headers follow a template at the org. and does not seem to have problem with the earlier versions of LO (v6.4) when opened and converted to pdf.

Is this a known issue? What should I do? manually fixing the header is out of question since there could be thousands of files.

[mikekaganski]

@omidghazara

Is this a known issue?

Who knows. It really doesn’t matter, because the best thing to do is always not to spend time finding out “if it’s known”, but just go ahead to the next point, which is

@omidghazara

What should I do?

File a bug report with a sample problematic file, and screenshots/PDFs of the problem, with clearly marked problematic spots.

By the way, you never mentioned the version that fails for you (7.x is not a version, at least two pieces are necessary), only the version where it worked - did you try the newest ones?

[omidghazara]

I meant I have tried several of the v7.x.y versions (I lost track of them ;)) including the latest ones (7.6.2, 7.5.7) as the first trials, since we already have v7.1 in prod. The only resort is a downgrade to v6.4.7 as of now for us I guess until a fix shows up. Thanks for the response.

[keme1]

Also, Microsoft Office and LibreOffice seem to adhere to different “schools of typography” where things like spacing and alignment are treated slightly differently. This difference comes into play frequently when there is a conversion involving marginalia, adjoining/overlapping “leading” regions, indents, etc.

Seeing that it is OK for you in older versions and you observe a regression now, this difference may or may not be part of the cause. Anyway, I thought it worth mentioning.

[mikekaganski]

Note that in importing/exporting, the difference in the concepts only means that the filters must do the correct calculations/transformations, to result in documents looking as close to the original as possible. If some visual result is possible in both, this must be achieved by the filters. Only when something is limited in the formats themselves, it is justifiable to have visual difference; all other cases are bugs.