as for the xml file, the problem started from the fact that I couldn’t send the file as an email from my work’s desktop to my mobile (it was rejected by exchange server - only mail I have access to - I’ve still haven’t figured out why… anyway…), so I opened it, I copied the content, I pasted it on github and then I opened my mobile browser and I got the text from github, somewhere on the whole process greek characters broke

I opened the file from the disk now and it is utf-8-bom, but it didn’t matter I guess

is it an xml file you can share, or does it contain private data ?
can you find out which greek letters will trigger this, by making an empty xml with just some greek words ?

in french it is triggered, for example, by a single word Mosaïque because of the ï.
@guy038 also found out, that if you combine the two words mosaïque étaitin a new file it will work correctly, but also était alone in a file will not work.

if it’s easy to reproduce, you could file an issue at github: https://github.com/notepad-plus-plus/notepad-plus-plus/issues.
(and then hope it gets looked at by the developers, due to over 2400 open issues at the moment)

i for myself use utf-8 file encoding only, and convert all files to utf-8 if they are not.

something that’s quick to try out if it works for you:
a user told us, that he converts all problematic documents with encoding problems to utf-8-bom (utf-8 with a byte order mark header), because the bom header will explicitly state which format this file has, and utf-8-bom seems to be compatible with all his applications and web services.

it would be interesting if your encoding issue gets better using bom, and what happens on android with a bom file.

I’ve got that one disabled a long time ago (it was suggested here to me)

yes, my apology, i forgot.

as I said and I’ll repeat myself, the bug is definitely in 7.6.3,
I replaced the executable and it worked fine, just fine

i will test that too, both 7.6.2 and 7.6.3 exe, with my files and hope it did not get worse in 7.6.3, or at least that we find a workaround.
good idea to use the old exe if it works for you.

as I said and I’ll repeat myself, the bug is definitely in 7.6.3,
I replaced the executable and it worked fine, just fine

unfortunately the current auto detection for character encodings is still broken in notepad++.
maybe that’s also the reason why your xml got saved with wrong characters, if you have it enabled.

for example if your xml was utf-8 in it’s original form and notepad++ detected it as another (wrong) encoding (example: encoding > character sets > vietnamese > windows-1258, like it happens many times for french and spanish)
if you now edit such a file and save it, the characters can get messed up.

currently, if you have it enabled, it’s recommended to disable “autodetect character encoding” in settings > preferences > misc, like seen at the screenshot below.
then restart notepad++ and retry with your xml file (i hope you still have an old backup of it, with correct greek characters).

hope it helps you a bit.