2019-08-23 05:17:15 +02:00
|
|
|
extern crate base64;
|
|
|
|
|
|
|
|
use self::base64::encode;
|
2019-10-12 11:32:59 +02:00
|
|
|
use http::retrieve_asset;
|
2019-09-29 23:15:49 +02:00
|
|
|
use regex::Regex;
|
2019-12-13 01:29:21 +01:00
|
|
|
use reqwest::Client;
|
2019-12-06 02:20:09 +01:00
|
|
|
use std::collections::HashMap;
|
2019-09-29 23:15:49 +02:00
|
|
|
use url::{ParseError, Url};
|
|
|
|
|
2019-12-09 18:41:21 +01:00
|
|
|
/// This monster of a regex is used to match any kind of URL found in CSS.
|
2019-12-06 20:59:13 +01:00
|
|
|
///
|
2019-12-09 18:41:21 +01:00
|
|
|
/// There are roughly three different categories that a found URL could fit
|
|
|
|
/// into:
|
2019-12-06 20:59:13 +01:00
|
|
|
/// - Font [found after a src: property in an @font-family rule]
|
|
|
|
/// - Stylesheet [denoted by an @import before the url
|
|
|
|
/// - Image [covers all other uses of the url() function]
|
|
|
|
///
|
|
|
|
/// This regex aims to extract the following information:
|
|
|
|
/// - What type of URL is it (font/image/css)
|
|
|
|
/// - Where is the part that needs to be replaced (incl any wrapping quotes)
|
|
|
|
/// - What is the URL (excl any wrapping quotes)
|
|
|
|
///
|
2019-12-09 18:41:21 +01:00
|
|
|
/// Essentially, the regex can be broken down into two parts:
|
2019-12-06 20:59:13 +01:00
|
|
|
///
|
|
|
|
/// `(?:(?P<import>@import)|(?P<font>src\s*:)\s+)?`
|
2019-12-09 18:41:21 +01:00
|
|
|
/// This matches the precursor to a font or CSS URL, and fills in a match under
|
|
|
|
/// either `<import>` (if it's a CSS URL) or `<font>` (if it's a font).
|
2019-12-06 20:59:13 +01:00
|
|
|
/// Determining whether or not it's an image can be done by the negation of both
|
2019-12-09 18:41:21 +01:00
|
|
|
/// of these. Either zero or one of these can match.
|
2019-12-06 20:59:13 +01:00
|
|
|
///
|
|
|
|
/// `url\((?P<to_repl>['"]?(?P<url>[^"'\)]+)['"]?)\)`
|
2019-12-09 18:41:21 +01:00
|
|
|
/// This matches the actual URL part of the url(), and must always match. It also
|
2019-12-06 20:59:13 +01:00
|
|
|
/// sets `<to_repl>` and `<url>` which correspond to everything within
|
|
|
|
/// `url(...)` and a usable URL, respectively.
|
|
|
|
///
|
|
|
|
/// Note, however, that this does not perform any validation of the found URL.
|
2019-12-09 18:41:21 +01:00
|
|
|
/// Malformed CSS could lead to an invalid URL being present. It is therefore
|
|
|
|
/// recomended that the URL gets manually validated.
|
2019-12-06 22:46:52 +01:00
|
|
|
const CSS_URL_REGEX_STR: &str = r###"(?:(?:(?P<stylesheet>@import)|(?P<font>src\s*:))\s+)?url\((?P<to_repl>['"]?(?P<url>[^"'\)]+)['"]?)\)"###;
|
2019-12-06 19:53:44 +01:00
|
|
|
|
2019-09-29 23:15:49 +02:00
|
|
|
lazy_static! {
|
|
|
|
static ref HAS_PROTOCOL: Regex = Regex::new(r"^[a-z0-9]+:").unwrap();
|
|
|
|
static ref REGEX_URL: Regex = Regex::new(r"^https?://").unwrap();
|
2019-12-06 19:53:44 +01:00
|
|
|
static ref REGEX_CSS_URL: Regex = Regex::new(CSS_URL_REGEX_STR).unwrap();
|
2019-09-29 23:15:49 +02:00
|
|
|
}
|
2019-08-24 00:48:08 +02:00
|
|
|
|
2019-10-10 15:23:00 +02:00
|
|
|
const MAGIC: [[&[u8]; 2]; 19] = [
|
2019-08-24 00:48:08 +02:00
|
|
|
// Image
|
|
|
|
[b"GIF87a", b"image/gif"],
|
|
|
|
[b"GIF89a", b"image/gif"],
|
|
|
|
[b"\xFF\xD8\xFF", b"image/jpeg"],
|
|
|
|
[b"\x89PNG\x0D\x0A\x1A\x0A", b"image/png"],
|
|
|
|
[b"<?xml ", b"image/svg+xml"],
|
|
|
|
[b"<svg ", b"image/svg+xml"],
|
|
|
|
[b"RIFF....WEBPVP8 ", b"image/webp"],
|
|
|
|
[b"\x00\x00\x01\x00", b"image/x-icon"],
|
|
|
|
// Audio
|
|
|
|
[b"ID3", b"audio/mpeg"],
|
|
|
|
[b"\xFF\x0E", b"audio/mpeg"],
|
|
|
|
[b"\xFF\x0F", b"audio/mpeg"],
|
|
|
|
[b"OggS", b"audio/ogg"],
|
|
|
|
[b"RIFF....WAVEfmt ", b"audio/wav"],
|
|
|
|
[b"fLaC", b"audio/x-flac"],
|
|
|
|
// Video
|
|
|
|
[b"RIFF....AVI LIST", b"video/avi"],
|
|
|
|
[b"....ftyp", b"video/mp4"],
|
|
|
|
[b"\x00\x00\x01\x0B", b"video/mpeg"],
|
|
|
|
[b"....moov", b"video/quicktime"],
|
|
|
|
[b"\x1A\x45\xDF\xA3", b"video/webm"],
|
|
|
|
];
|
2019-08-23 05:17:15 +02:00
|
|
|
|
|
|
|
pub fn data_to_dataurl(mime: &str, data: &[u8]) -> String {
|
2019-10-10 15:23:00 +02:00
|
|
|
let mimetype = if mime.is_empty() {
|
2019-08-23 20:24:45 +02:00
|
|
|
detect_mimetype(data)
|
2019-08-23 05:17:15 +02:00
|
|
|
} else {
|
2019-08-23 20:24:45 +02:00
|
|
|
mime.to_string()
|
|
|
|
};
|
2019-08-23 05:17:15 +02:00
|
|
|
format!("data:{};base64,{}", mimetype, encode(data))
|
|
|
|
}
|
|
|
|
|
2019-09-29 23:15:49 +02:00
|
|
|
pub fn detect_mimetype(data: &[u8]) -> String {
|
2019-09-22 02:06:00 +02:00
|
|
|
for item in MAGIC.iter() {
|
2019-08-24 00:48:08 +02:00
|
|
|
if data.starts_with(item[0]) {
|
2019-10-10 15:23:00 +02:00
|
|
|
return String::from_utf8(item[1].to_vec()).unwrap();
|
2019-08-24 00:48:08 +02:00
|
|
|
}
|
|
|
|
}
|
2019-10-10 15:23:00 +02:00
|
|
|
"".to_owned()
|
2019-08-23 05:17:15 +02:00
|
|
|
}
|
|
|
|
|
2019-10-10 15:23:00 +02:00
|
|
|
pub fn url_has_protocol<T: AsRef<str>>(url: T) -> bool {
|
|
|
|
HAS_PROTOCOL.is_match(url.as_ref().to_lowercase().as_str())
|
2019-09-29 23:15:49 +02:00
|
|
|
}
|
2019-08-23 05:17:15 +02:00
|
|
|
|
2019-10-10 15:23:00 +02:00
|
|
|
pub fn is_data_url<T: AsRef<str>>(url: T) -> Result<bool, ParseError> {
|
|
|
|
Url::parse(url.as_ref()).and_then(|u| Ok(u.scheme() == "data"))
|
2019-09-29 23:15:49 +02:00
|
|
|
}
|
2019-08-24 00:48:08 +02:00
|
|
|
|
2019-10-10 15:23:00 +02:00
|
|
|
pub fn is_valid_url<T: AsRef<str>>(path: T) -> bool {
|
|
|
|
REGEX_URL.is_match(path.as_ref())
|
2019-09-29 23:15:49 +02:00
|
|
|
}
|
|
|
|
|
2019-10-10 15:23:00 +02:00
|
|
|
pub fn resolve_url<T: AsRef<str>, U: AsRef<str>>(from: T, to: U) -> Result<String, ParseError> {
|
|
|
|
let result = if is_valid_url(to.as_ref()) {
|
|
|
|
to.as_ref().to_string()
|
2019-09-29 23:15:49 +02:00
|
|
|
} else {
|
2019-10-10 15:23:00 +02:00
|
|
|
Url::parse(from.as_ref())?
|
|
|
|
.join(to.as_ref())?
|
|
|
|
.as_ref()
|
|
|
|
.to_string()
|
2019-09-29 23:15:49 +02:00
|
|
|
};
|
|
|
|
Ok(result)
|
2019-08-23 05:17:15 +02:00
|
|
|
}
|
2019-10-12 11:32:59 +02:00
|
|
|
|
2019-10-12 13:05:07 +02:00
|
|
|
pub fn resolve_css_imports(
|
2019-12-06 01:01:03 +01:00
|
|
|
cache: &mut HashMap<String, String>,
|
2019-12-13 01:29:21 +01:00
|
|
|
client: &Client,
|
2019-10-12 13:05:07 +02:00
|
|
|
css_string: &str,
|
2019-12-06 02:05:52 +01:00
|
|
|
as_dataurl: bool,
|
2019-10-12 13:05:07 +02:00
|
|
|
href: &str,
|
2019-12-06 20:59:13 +01:00
|
|
|
opt_no_images: bool,
|
2019-10-12 13:05:07 +02:00
|
|
|
opt_silent: bool,
|
2019-12-06 00:15:06 +01:00
|
|
|
) -> String {
|
2019-10-12 11:32:59 +02:00
|
|
|
let mut resolved_css = String::from(css_string);
|
2019-10-12 13:05:07 +02:00
|
|
|
|
2019-12-06 19:53:44 +01:00
|
|
|
for link in REGEX_CSS_URL.captures_iter(&css_string) {
|
2019-12-06 00:15:06 +01:00
|
|
|
let target_link = link.name("url").unwrap().as_str();
|
2019-12-05 23:42:07 +01:00
|
|
|
|
2019-12-06 20:59:13 +01:00
|
|
|
// Determine the type of link
|
|
|
|
let is_stylesheet = link.name("stylesheet").is_some();
|
|
|
|
let is_font = link.name("font").is_some();
|
|
|
|
let is_image = !is_stylesheet && !is_font;
|
|
|
|
|
2019-12-05 23:42:07 +01:00
|
|
|
// Generate absolute URL for content
|
|
|
|
let embedded_url = match resolve_url(href, target_link) {
|
|
|
|
Ok(url) => url,
|
|
|
|
Err(_) => continue, // Malformed URL
|
2019-10-12 13:05:07 +02:00
|
|
|
};
|
|
|
|
|
2019-12-09 18:41:21 +01:00
|
|
|
// Download the asset. If it's more CSS, resolve that too
|
2019-12-06 20:59:13 +01:00
|
|
|
let content = if is_stylesheet {
|
2019-12-06 00:15:06 +01:00
|
|
|
// The link is an @import link
|
2019-12-06 20:59:13 +01:00
|
|
|
retrieve_asset(
|
2019-12-06 02:41:43 +01:00
|
|
|
cache,
|
2019-12-13 01:29:21 +01:00
|
|
|
client,
|
2019-12-06 02:41:43 +01:00
|
|
|
&embedded_url,
|
|
|
|
false, // Formating as data URL will be done later
|
|
|
|
"text/css", // Expect CSS
|
|
|
|
opt_silent,
|
|
|
|
)
|
|
|
|
.map(|(content, _)| {
|
|
|
|
resolve_css_imports(
|
|
|
|
cache,
|
2019-12-13 01:29:21 +01:00
|
|
|
client,
|
2019-12-06 02:41:43 +01:00
|
|
|
&content,
|
2019-12-09 18:41:21 +01:00
|
|
|
true, // Finally, convert to a dataurl
|
2019-12-06 02:41:43 +01:00
|
|
|
&embedded_url,
|
2019-12-06 20:59:13 +01:00
|
|
|
opt_no_images,
|
2019-12-06 02:41:43 +01:00
|
|
|
opt_silent,
|
|
|
|
)
|
2019-12-06 20:59:13 +01:00
|
|
|
})
|
|
|
|
} else if (is_image && !opt_no_images) || is_font {
|
2019-12-06 00:15:06 +01:00
|
|
|
// The link is some other, non-@import link
|
2019-12-06 20:59:13 +01:00
|
|
|
retrieve_asset(
|
2019-12-06 02:41:43 +01:00
|
|
|
cache,
|
2019-12-13 01:29:21 +01:00
|
|
|
client,
|
2019-12-06 02:41:43 +01:00
|
|
|
&embedded_url,
|
|
|
|
true, // Format as data URL
|
|
|
|
"", // Unknown MIME type
|
|
|
|
opt_silent,
|
|
|
|
)
|
2019-12-06 20:59:13 +01:00
|
|
|
.map(|(a, _)| a)
|
|
|
|
} else {
|
|
|
|
// If it's a datatype that has been opt_no'd out of, replace with
|
|
|
|
// absolute URL
|
|
|
|
|
|
|
|
Ok(embedded_url.clone())
|
2019-12-06 02:41:43 +01:00
|
|
|
}
|
|
|
|
.unwrap_or_else(|e| {
|
2019-12-09 18:41:21 +01:00
|
|
|
eprintln!("Warning: {}", e);
|
2019-12-06 00:37:37 +01:00
|
|
|
|
2019-12-09 18:41:21 +01:00
|
|
|
// If failed to resolve, replace with absolute URL
|
2019-12-06 00:37:37 +01:00
|
|
|
embedded_url
|
|
|
|
});
|
2019-12-06 00:15:06 +01:00
|
|
|
|
|
|
|
let replacement = format!("\"{}\"", &content);
|
2019-12-06 17:37:05 +01:00
|
|
|
let dest = link.name("to_repl").unwrap();
|
2019-12-06 19:52:51 +01:00
|
|
|
let offset = resolved_css.len() - css_string.len();
|
|
|
|
let target_range = (dest.start() + offset)..(dest.end() + offset);
|
2019-12-06 00:15:06 +01:00
|
|
|
|
2019-12-06 19:52:51 +01:00
|
|
|
resolved_css.replace_range(target_range, &replacement);
|
2019-10-12 11:32:59 +02:00
|
|
|
}
|
|
|
|
|
2019-12-06 02:05:52 +01:00
|
|
|
if as_dataurl {
|
|
|
|
data_to_dataurl("text/css", resolved_css.as_bytes())
|
|
|
|
} else {
|
|
|
|
resolved_css
|
|
|
|
}
|
2019-10-12 13:05:07 +02:00
|
|
|
}
|
2019-12-12 03:13:11 +01:00
|
|
|
|
|
|
|
pub fn clean_url<T: AsRef<str>>(url: T) -> String {
|
|
|
|
let mut result = Url::parse(url.as_ref()).unwrap();
|
|
|
|
// Clear fragment
|
|
|
|
result.set_fragment(None);
|
|
|
|
// Get rid of stray question mark
|
|
|
|
if result.query() == Some("") {
|
|
|
|
result.set_query(None);
|
|
|
|
}
|
|
|
|
result.to_string()
|
|
|
|
}
|