Can you reproduce this on https://try.gogs.io ?

Yeah! Here is! 👍

Thanks! How does GitHub handles it?

I think this is expected behavior. Gogs does not give any assumption for raw format of content, and browser determines it automatically.

Since you use .html as extension, it is reasonable that browser treats it as an HTML page.

Yeah, I thought same, but GitHub use the <pre> tag.

For example, a web page in RAW mode in GitHub

but GitHub use the

 tag.

How do you know?

Thanks to inspect element of Chrome...

So, the HTML text isn't interpreted by browser 👍

I see. I think GitHub only does this for .html and .xml?

I think that for all of them. I'm just see that javascript and php text use it too.

OK... I think it involves something else bigger... I believe if you do curl xxx, GitHub responses pure raw content of any types.

It's possible!

how about this:

func ServeData(ctx *middleware.Context, name string, reader io.Reader) error {
    buf := make([]byte, 1024)
    n, _ := reader.Read(buf)
    if n > 0 {
        buf = buf[:n]
    }

    _, isTextFile := base.IsTextFile(buf)
    if !isTextFile {
        _, isImageFile := base.IsImageFile(buf)
        if !isImageFile {
            ctx.Resp.Header().Set("Content-Disposition", "attachment; filename="+path.Base(ctx.Repo.TreeName))
            ctx.Resp.Header().Set("Content-Transfer-Encoding", "binary")
        }
    } else {
        ctx.Resp.Header().Set("Content-Type", "text/plain; charset=utf-8") // <<<<<<
    }
    ctx.Resp.Write(buf)
    _, err := io.Copy(ctx.Resp, reader)
    return err
}

@xxxTonixxx I think it is a feature of the browser(FF/IE/Chrome), not github

@novaeye Hmmm, It's true that in source code doesn't appear the <pre> tag ... Good catch! 👍

Thanks @novaeye !

@xxxTonixxx here we go: https://try.gogs.io/xxxtonixxx/bug_html_rendering_raw_mode/raw/master/toni/LenMar/HTML/index.html

It's ok! but... There is a problem with encoding characters (áéíóú...).

• These show so:
  
• And these must show so:
  

Github raw view serves files with Content-Type: text. This means that neither HTML, CSS or JS can be used directly in web pages.

There's a services that serve files from Github with proper content type headers: https://rawgit.com

Maybe in Gogs we could have a URL parameter to it: http://.../index.html?raw=1. So people could still use files directly if they want.

(I think it should default to false if not provided so something like <style href="my.css"> will work by default in a web page, without having to add ?raw=0).

That will be interesnting 👍, but... to work from the address bar doesn't think that it's the best way...

@xxxTonixxx your screenshots more like browser auto-detection thing for me. My Chrome show what I expect:

Ok! I already have changed and right now it's working... but, it's strange that with GitHub's raw mode was working without I change nothing... With Firefox was equal but with IE is working good!

Thank you!!! 😄 👍

@xxxTonixxx GitHub may do encoding transformation, but Gogs serves "raw" from Git blob.

what if somebody want a raw html rendered.

To get something like gitpages, or so.

Use case:
I want to have a tiddlywiky in my repo, and to serve it directly from gogs.

@pykiss Hmm... maybe you can file another issue and ask likes Allow request header to set HTML rendered in raw mode?

@pykiss have you already filed a new issue? I would also have such a feature back. All my analysis in R (statistics) are rendered in the end to a html page. Normally I just compare different reports by looking in my git history but now it is broken. It worked until last release.

@tkschmidt

Normally I just compare different reports by looking in my git history but now it is broken. It worked until last release.

What does this mean specifically?

sry, for bad wording. Until last release while looking at the raw file type i got my rendered html. Now, because it is wrapped into another html it is broken and I get the view above.

@tkschmidt OK.. I think this is intended to not render.

You may want to follow up on #2593

Still unsolved?

I wget a python file via its raw url, but the downloaded file is in html format ???????