Greek characters turned into question mark in a docx file

csm · February 6, 2020, 5:34pm

Hello,

I used to write my text (in greek) since 2014 using Office 2010 in this file.

Recently I decided to install Libreoffice (6.3.4) as Office 2010 is getting out of support this year.

Yesterday I wrote a few paragraph as usual, then saved the file.

When I reopened it today, almost all the greek caracters (not the latin one through) were turned into ‘???’.

The file is here : Filebin | a1yhpv4nlbt07h0j

I made a copy of the file and tried to use different tools to repair it, but nothing worked… did I just lost 6 years of work as I don’t have a backup ?

Thanks for any help you can provide.

anon73440385 · February 6, 2020, 5:41pm

did I just lost 6 years of work as I don’t have a backup

This makes my words fail …

The file you have provided above is no .docx file any more and there is mostly hex 3F - so no encoding problem but a complete corruption by whatever content.

csm · February 6, 2020, 5:52pm

How could that happen ?

I mean, I really only did what I said with this file…

anon73440385 · February 6, 2020, 6:12pm

Check your system for a filesystem corruption - may be your disk is going to die, may be there’s really a bug, may be there was someone/something overwriting this by accident. Can only make more or less educated guesses from the distance.

mikekaganski · February 7, 2020, 7:18am

This does not look like a filesystem corruption. The file looks “valid” - in the sense that it’s like it was purposefully saved as “text with encoding” with e.g. Windows-1252 encoding, or just as plain text on a non-Greek system (so that its default encoding does not contain those characters). That process always converts missing characters into question marks. And of course, that is always accompanied with a warning like “saving as text may loose …”.

No idea how OP could choose the file type without knowing that - but I am 100% sure that text filter was used in saving. Might it be that the file was plain text with some encoding from the start, and was authored like that all way back, and now just a wrong encoding was used when saving? Just a guess.