Changed the UTF Encoding instance to avoid inserting the BOM character

[chrismld]

Fix to solve issue #9

[chrismld]

@nblumhardt it seems that this won't be possible for now, look https://github.com/dotnet/coreclr/issues/5000

[nblumhardt]

Thanks Christian!

I don't think that the linked issue is the problem here, the compiler error is:

C:\projects\serilog-sinks-file\src\Serilog.Sinks.File\Sinks\File\FileSink.cs(59,71): error CS0246: The type or namespace name 'UTF8Encoding' could not be found (are you missing a using directive or an assembly reference?)

If the constructor accepting bool isn't present, the issue suggests that new UTF8Encoding() will do as we desire - a conditional compilation directive would cover both platforms in that case.

[chrismld]

@nblumhardt still not working, that constructor doesn't exist anymore

[nblumhardt]

The package:

        "System.Text.Encoding.Extensions": "4.0.11"

needs to be added to the netstandard1.3 dependency group for the UTF8Encoding type to appear.

Just a thought, shall we write it as:

new UTF8Encoding(encoderShouldEmitUTF8Identifier: false)

I.e., name the parameter so that there's absolutely no ambiguity?

The above works without any conditional compilation, so seems like the way to go 👍

[chrismld]

@nblumhardt not sure what happened but yesterday I just tried the same thing and it didn't work (guess it was my workstation). Anyways, I already submitted the change here, I'm gonna do it for the RollingFile project :)

[nblumhardt]

Thanks @christianhxc, great stuff :-)