@freezestudio can you attach the file please?

GitHub Setup error log file:
this rename .log to .txt:
install.txt

The file is containing null bytes which we then detect as binary type. However the file seems to use UTF 16 as encoding. We only support that if the file has a BOM in the beginning. I think other editors just try to open the file with UTF 16 if they detect null bytes but this can also fail.

I think we should allow to change the encoding from the editor in all cases even if we think the file is binary.

This one feature prevents us from using VS Code here at work. We often have Markdown files with embedded null bytes (and other non-printable characters).

@iffy may I ask why your markdown files contain null bytes?

Ok, I think having a way to force open a file that we think is binary is still valid. E.g. similar how for large images we ask you to open them externally there could be a link to force open them inside the editor:

I would accept a PR on such a solution 👍

I tried... I changed it to add a button labeled "Open anyway" but since this code base is enormous and unfamiliar, I have no idea what function to call to open a file, let alone with an explicit encoding :) Somewhere down the openFileFolderPicker -> open(IOpenConfiguration) chain....

Yeah sorry, the code is quite complex. You could start to debug through how editors open e.g. by adding a breakpoint to editorPart.openEditor(). I think for this to work we need a way to "force" open a binary file as text, even if our check returns that it is a binary file (here).

@bpasero what about a couple lines up options.acceptTextOnly? How can we turn that off?

There should probably be a way to control this when opening that file to override.

Could we just have a setting to always open whatever we tell it to open? Don't worry about protecting me from myself... ;)

The file is containing null bytes which we then detect as binary type. However the file seems to use UTF 16 as encoding. We only support that if the file has a BOM in the beginning. I think other editors just try to open the file with UTF 16 if they detect null bytes but this can also fail.

That's insane and has to be fixed. I just tried to open a couple of files WER generated and... no go.

Not only it is unreasonable not to open UTF-16 encoded text files without a BOM since there too many of those files out there, it is completely ridiculous when Microsoft itself generates such files all over the place.

I think we should allow to change the encoding from the editor in all cases even if we think the file is binary.

I agree.

But... Two years and still nothing? ☹️ I don't blame you personally, Benjamin, but I certainly blame Microsoft.

-"The file will not be displayed because is it either binary, very large or an unsupported text encoding."
-"WAT? That's UTF-16!"
-"Sorry, bro. I only support ASCII and EBCDIC."

That's how it sounds.

Being able to open text files is the first requirement from a source code editor, or an advanced text editor, or any text editor...

There is now a new action to force open a file that we think is binary into the editor. You can then change the encoding accordingly.