Yes. In general, WinSCP first tries to load the file using the default encoding and if that fails, it falls back to the other. But the problem is that any file can be loaded with Ansi encoding, so if you have Ansi set as the default, fall back to UTF-8 cannot happen. But if you have UTF-8 encoding set as the default, WinSCP can fallback to Ansi, if the file is not valid UTF-8 file.

I'm not clear on "encoding autodetection works fine, when the default encoding is configured as UTF-8." You mean when UTF-8 is default, it accurately detects non-UTF-8 encoded files?

I may or may not be opening a UTF-8 file (and there is no way to tell before opening), but I went ahead and set to default as I believe that will be most common for me.

After some tests, I believe the encoding autodetection works fine, when the default encoding is configured as UTF-8. So all I've done for now is to default to UTF-8 since the next major release.
For you, it should be enough, if you configure the UTF-8 as default manually in the current version of WinSCP. Let me know.

Thank you, Martin. Even a way to know the encoding of a file would be very helpful.

I'm not aware of any relevant change. If you believe otherwise, can you please try to find out in version of WinSCP changed the behavior?
https://sourceforge.net/projects/winscp/files/WinSCP/

---

I have added a request for encoding detection improvement to the tracker:
Issue 2397 – Default to UTF-8 encoding in internal editor
You can vote for it there.

Yes, I already stated the files were UTF-8. Behavior seems to have recently changed; that is why I reported this issue.

I don't know which file (of all the files that I edit) is what encoding. How can I know before editing? Why doesn't the editor detect and set appropriately?

The file has UTF-8 encoding.
So you need to use that when editing the file.
Once you have modified and saved the file using different encoding (Ansi), it becomes corrupted and trying to reload it with UTF-8 won't help.
If all your files are UTF-8, please configure it as the Default encoding in preferences.

Default encoding is 1252 ANSI Latin I. Attached file is before uploaded to server - okay?

The diamonds usually show up after I edit the character before or after one of these: “ or †(double quotes).

Thanks for your report. Can you post an example file? Downloaded (not edited) in binary mode from the server?

Also, what Default encoding do you have configured here?
https://winscp.net/eng/docs/ui_pref_editor_internal

And, if you try to change the encoding after saving in the editor, you get:

---------------------------
Error
---------------------------
Error loading file 'C:\Users\justin\AppData\Local\Temp\scp51482\home3\bereanb2\public_html\1Corinthians\notes\609-610_1Corinthians_7-27-25.html' using 'UTF-8' encoding.
---------------------------
OK Help
---------------------------

I have been editing HTML UTF-8 files from the remote server for a long time and had no issue until recently, like v6.5.0/2 time frame. Now when I open the file, make a modification, save, reload in the browser, I get ?-diamonds around the edit. So I suspect there is a regression in properly setting encoding type in the editor.

The editor shows under Encoding "1252". Since I didn't have a problem before, I never checked this until now, so I can't say what it said before.

Windows 10, files created via Save As HTML from LibreOffice Writer.