Deliver text files, JSON, XML and JavaScript with `charset=utf-8`

[thekid]

Motivation

Why is that important?

Even though browsers sometimes ignore the value of the Content-Type header and try to sniff the content (see also: X-Content-Type-Options hint), it’s indicated to always send the appropriate media type and charset for the response as, among other:

• the media type defines both the data format and how that data is intended to be processed by browsers
• not sending the appropriate charset, where appropriate, may prevent things from being rendered correctly thus creating a bad user experience (see also: meta-charset-utf-8 hint)
• javascript resources served with the wrong media type may be blocked

See https://webhint.io/docs/user-guide/hints/hint-content-type/

[thekid]

Views are already sent with a charset in the content type, see here:

https://github.com/xp-forge/frontend/blob/a6349228938a2d80357d33337606b5c10d5ea0fd/src/main/php/web/frontend/View.class.php#L118

However, we could use xp::ENCODING here instead of hard-wiring it.

[thekid]

https://webhint.io/docs/user-guide/server-configurations/apache/ lists the following file extensions for adding utf-8 to:

Extension	util.MimeType::getByFileName()	Action
.appcache	text/cache-manifest	text/*
.atom	application/atom+xml	Listed
.css	text/css	text/*
.js	application/javascript	Listed
.json	application/json	Listed
.manifest	text/cache-manifest	text/*
.map	application/json	Listed
.mjs	application/javascript	Listed
.rdf	application/rdf+xml	Listed
.rss	application/rss+xml	Listed
.vtt	text/vtt	text/*
.webmanifest	application/manifest+json	Listed
.xml	application/xml	Listed

https://webhint.io/docs/user-guide/server-configurations/iis/ includes image/svg+xml to include in this list.

We could also extend the rule to include application/*+xml and application/*+json, which would make the rule more generic.

[thekid]

I'm a bit reluctant to automatically set a charset parameter for XML documents, as they typically start with this line:

<?xml version="1.0" encoding="UTF-8"?>

...and this may create the situation where our Content-Type header says one thing and the XML file says something different. On the other hand, no one in his or her right mind would deliver XML files with anything other than utf-8 today, would they?

See also https://stackoverflow.com/questions/3272534/what-content-type-value-should-i-send-for-my-xml-sitemap

[thekid]

Released in https://github.com/xp-forge/frontend/releases/tag/v3.3.0