Merge pull request #167 from snshn/version-bump

version bump

.appveyor.yml

@@ -1,131 +0,0 @@
-# Appveyor configuration template for Rust using rustup for Rust installation
-# https://github.com/starkat99/appveyor-rust
-
-## Operating System (VM environment) ##
-
-# Rust needs at least Visual Studio 2013 Appveyor OS for MSVC targets.
-os: Visual Studio 2015
-
-## Build Matrix ##
-
-# This configuration will setup a build for each channel & target combination (12 windows
-# combinations in all).
-#
-# There are 3 channels: stable, beta, and nightly.
-#
-# Alternatively, the full version may be specified for the channel to build using that specific
-# version (e.g. channel: 1.5.0)
-#
-# The values for target are the set of windows Rust build targets. Each value is of the form
-#
-# ARCH-pc-windows-TOOLCHAIN
-#
-# Where ARCH is the target architecture, either x86_64 or i686, and TOOLCHAIN is the linker
-# toolchain to use, either msvc or gnu. See https://www.rust-lang.org/downloads.html#win-foot for
-# a description of the toolchain differences.
-# See https://github.com/rust-lang-nursery/rustup.rs/#toolchain-specification for description of
-# toolchains and host triples.
-#
-# Comment out channel/target combos you do not wish to build in CI.
-#
-# You may use the `cargoflags` and `RUSTFLAGS` variables to set additional flags for cargo commands
-# and rustc, respectively. For instance, you can uncomment the cargoflags lines in the nightly
-# channels to enable unstable features when building for nightly. Or you could add additional
-# matrix entries to test different combinations of features.
-environment:
-  matrix:
-
-### MSVC Toolchains ###
-
-  # Stable 64-bit MSVC
-    - channel: stable
-      target: x86_64-pc-windows-msvc
-  # Stable 32-bit MSVC
-    - channel: stable
-      target: i686-pc-windows-msvc
-  # Beta 64-bit MSVC
-    - channel: beta
-      target: x86_64-pc-windows-msvc
-  # Beta 32-bit MSVC
-    - channel: beta
-      target: i686-pc-windows-msvc
-  # Nightly 64-bit MSVC
-    - channel: nightly
-      target: x86_64-pc-windows-msvc
-      cargoflags: --features "unstable"
-  # Nightly 32-bit MSVC
-    - channel: nightly
-      target: i686-pc-windows-msvc
-      cargoflags: --features "unstable"
-
-### GNU Toolchains ###
-
-  # Stable 64-bit GNU
-    - channel: stable
-      target: x86_64-pc-windows-gnu
-      MINGW_PATH: 'C:\mingw-w64\x86_64-6.3.0-posix-seh-rt_v5-rev1\mingw64\bin'
-  # Stable 32-bit GNU
-    - channel: stable
-      target: i686-pc-windows-gnu
-      MINGW_PATH: 'C:\MinGW\bin'
-  # Beta 64-bit GNU
-    - channel: beta
-      target: x86_64-pc-windows-gnu
-      MINGW_PATH: 'C:\mingw-w64\x86_64-6.3.0-posix-seh-rt_v5-rev1\mingw64\bin'
-  # Beta 32-bit GNU
-    - channel: beta
-      target: i686-pc-windows-gnu
-      MINGW_PATH: 'C:\MinGW\bin'
-  # Nightly 64-bit GNU
-    - channel: nightly
-      target: x86_64-pc-windows-gnu
-      MINGW_PATH: 'C:\mingw-w64\x86_64-6.3.0-posix-seh-rt_v5-rev1\mingw64\bin'
-      cargoflags: --features "unstable"
-  # Nightly 32-bit GNU
-    - channel: nightly
-      target: i686-pc-windows-gnu
-      MINGW_PATH: 'C:\MinGW\bin'
-      cargoflags: --features "unstable"
-
-### Allowed failures ###
-
-# See Appveyor documentation for specific details. In short, place any channel or targets you wish
-# to allow build failures on (usually nightly at least is a wise choice). This will prevent a build
-# or test failure in the matching channels/targets from failing the entire build.
-matrix:
-  allow_failures:
-    - channel: nightly
-    - channel: beta
-
-# If you only care about stable channel build failures, uncomment the following line:
-    #- channel: beta
-
-## Install Script ##
-
-# This is the most important part of the Appveyor configuration. This installs the version of Rust
-# specified by the 'channel' and 'target' environment variables from the build matrix. This uses
-# rustup to install Rust.
-#
-# For simple configurations, instead of using the build matrix, you can simply set the
-# default-toolchain and default-host manually here.
-install:
-  - appveyor DownloadFile https://win.rustup.rs/ -FileName rustup-init.exe
-  - rustup-init -yv --default-toolchain %channel% --default-host %target%
-  - set PATH=%PATH%;%USERPROFILE%\.cargo\bin
-  - if defined MINGW_PATH set PATH=%PATH%;%MINGW_PATH%
-  - rustc -vV
-  - cargo -vV
-  - rustup component add rustfmt
-
-## Build Script ##
-
-# 'cargo test' takes care of building for us, so disable Appveyor's build stage. This prevents
-# the "directory does not contain a project or solution file" error.
-build: false
-
-# Uses 'cargo test' to run tests and build. Alternatively, the project may call compiled programs
-# directly or perform other testing commands. Rust will automatically be placed in the PATH
-# environment variable.
-test_script:
-  - cargo build --all --locked --verbose %cargoflags%
-  - cargo test --all --locked --verbose %cargoflags%

.github/workflows/build_gnu_linux.yml

@@ -0,0 +1,22 @@
+name: GNU/Linux
+
+on:
+  push:
+    branches: [ master ]
+
+jobs:
+  build:
+
+    strategy:
+      matrix:
+        os:
+          - ubuntu-latest
+        rust:
+          - stable
+    runs-on: ${{ matrix.os }}
+
+    steps:
+    - run: git config --global core.autocrlf false
+    - uses: actions/checkout@v2
+    - name: Build
+      run: cargo build --all --locked --verbose

.github/workflows/build_macos.yml

@@ -0,0 +1,22 @@
+name: macOS
+
+on:
+  push:
+    branches: [ master ]
+
+jobs:
+  build:
+
+    strategy:
+      matrix:
+        os:
+          - macos-latest
+        rust:
+          - stable
+    runs-on: ${{ matrix.os }}
+
+    steps:
+    - run: git config --global core.autocrlf false
+    - uses: actions/checkout@v2
+    - name: Build
+      run: cargo build --all --locked --verbose

.github/workflows/build_windows.yml

@@ -0,0 +1,22 @@
+name: Windows
+
+on:
+  push:
+    branches: [ master ]
+
+jobs:
+  build:
+
+    strategy:
+      matrix:
+        os:
+          - windows-latest
+        rust:
+          - stable
+    runs-on: ${{ matrix.os }}
+
+    steps:
+    - run: git config --global core.autocrlf false
+    - uses: actions/checkout@v2
+    - name: Build
+      run: cargo build --all --locked --verbose

.github/workflows/cd.yml

@@ -0,0 +1,54 @@
+# CD GitHub Actions workflow for Monolith
+
+name: CD
+
+on:
+  release:
+    types:
+    - created
+
+jobs:
+
+  windows:
+    runs-on: windows-2019
+    steps:
+    - run: git config --global core.autocrlf false
+    - name: Checkout the repository
+      uses: actions/checkout@v2
+    - name: Build and install the executable
+      run: cargo build --release
+    - uses: Shopify/upload-to-release@1.0.0
+      with:
+        name: monolith.exe
+        path: target\release\monolith.exe
+        repo-token: ${{ secrets.GITHUB_TOKEN }}
+
+  gnu_linux_armhf:
+    runs-on: ubuntu-18.04
+    steps:
+    - name: Checkout the repository
+      uses: actions/checkout@v2
+    - name: Prepare cross-platform environment
+      run: |
+        sudo mkdir -p /cross-build-arm
+        sudo echo "deb [arch=armhf] http://ports.ubuntu.com/ubuntu-ports/ bionic main" >> /etc/apt/sources.list
+        sudo apt-get update
+        sudo apt-get install -y gcc-arm-linux-gnueabihf libc6-armhf-cross libc6-dev-armhf-cross
+        sudo apt-get download libssl1.1:armhf libssl-dev:armhf
+        sudo dpkg -x libssl1.1*.deb /cross-build-arm
+        sudo dpkg -x libssl-dev*.deb /cross-build-arm
+        rustup target add arm-unknown-linux-gnueabihf
+        echo "::set-env name=C_INCLUDE_PATH::/cross-build-arm/usr/include"
+        echo "::set-env name=OPENSSL_INCLUDE_DIR::/cross-build-arm/usr/include/arm-linux-gnueabihf"
+        echo "::set-env name=OPENSSL_LIB_DIR::/cross-build-arm/usr/lib/arm-linux-gnueabihf"
+        echo "::set-env name=PKG_CONFIG_ALLOW_CROSS::1"
+        echo "::set-env name=RUSTFLAGS::-C linker=arm-linux-gnueabihf-gcc -L/usr/arm-linux-gnueabihf/lib -L/cross-build-arm/usr/lib/arm-linux-gnueabihf -L/cross-build-arm/lib/arm-linux-gnueabihf"
+    - name: Build the executable
+      run: |
+        cargo build --release --target=arm-unknown-linux-gnueabihf
+    - name: Attach artifact to the release
+      uses: Shopify/upload-to-release@1.0.0
+      with:
+        name: monolith-gnu-linux-armhf
+        path: target/arm-unknown-linux-gnueabihf/release/monolith
+        repo-token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/ci.yml

@@ -0,0 +1,30 @@
+name: CI
+
+on:
+  pull_request:
+    branches: [ master ]
+
+jobs:
+  build_and_test:
+
+    strategy:
+      matrix:
+        os:
+          - ubuntu-latest
+          - macos-latest
+          - windows-latest
+        rust:
+          - stable
+          - beta
+          - nightly
+    runs-on: ${{ matrix.os }}
+
+    steps:
+    - run: git config --global core.autocrlf false
+    - uses: actions/checkout@v2
+    - name: Build
+      run: cargo build --all --locked --verbose
+    - name: Run tests
+      run: cargo test --all --locked --verbose
+    - name: Check code formatting
+      run: cargo fmt --all -- --check

.travis.yml

@@ -1,37 +0,0 @@
-language: rust
-cache: cargo
-
-os:
-- linux
-- osx
-- windows
-
-rust:
-- stable
-- beta
-- nightly
-
-git:
-  autocrlf: false # don't mangle LF into CRLF on windows
-
-before_script:
-- rustup component add rustfmt
-
-script:
-- cargo build --all --locked --verbose
-- cargo test --all --locked --verbose
-- |
-  if [[ "$TRAVIS_RUST_VERSION" == "stable" ]]; then
-    cargo fmt --all -- --check
-  fi
-
-after_success: |
-  if [ "${TRAVIS_OS_NAME}" == linux ] && [ "${TRAVIS_RUST_VERSION}" == stable ]; then
-    docker build -t monolith .
-    docker run monolith monolith -V
-  fi
-
-jobs:
-  allow_failures:
-  - rust: nightly
-  fast_finish: true

Cargo.lock

@@ -61,6 +61,11 @@ name = "base64"
 version = "0.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 
+[[package]]
+name = "base64"
+version = "0.12.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+
 [[package]]
 name = "bitflags"
 version = "1.2.1"
@@ -568,10 +573,10 @@ dependencies = [
 
 [[package]]
 name = "monolith"
-version = "2.2.1"
+version = "2.2.4"
 dependencies = [
  "assert_cmd 0.12.0 (registry+https://github.com/rust-lang/crates.io-index)",
- "base64 0.11.0 (registry+https://github.com/rust-lang/crates.io-index)",
+ "base64 0.12.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "clap 2.33.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "cssparser 0.27.2 (registry+https://github.com/rust-lang/crates.io-index)",
  "html5ever 0.24.1 (registry+https://github.com/rust-lang/crates.io-index)",
@@ -1501,6 +1506,7 @@ dependencies = [
 "checksum atty 0.2.13 (registry+https://github.com/rust-lang/crates.io-index)" = "1803c647a3ec87095e7ae7acfca019e98de5ec9a7d01343f611cf3152ed71a90"
 "checksum autocfg 0.1.7 (registry+https://github.com/rust-lang/crates.io-index)" = "1d49d90015b3c36167a20fe2810c5cd875ad504b39cff3d4eae7977e6b7c1cb2"
 "checksum base64 0.11.0 (registry+https://github.com/rust-lang/crates.io-index)" = "b41b7ea54a0c9d92199de89e20e58d49f02f8e699814ef3fdf266f6f748d15c7"
+"checksum base64 0.12.0 (registry+https://github.com/rust-lang/crates.io-index)" = "7d5ca2cd0adc3f48f9e9ea5a6bbdf9ccc0bfade884847e484d452414c7ccffb3"
 "checksum bitflags 1.2.1 (registry+https://github.com/rust-lang/crates.io-index)" = "cf1de2fe8c75bc145a2f577add951f8134889b4795d47466a54a5c846d691693"
 "checksum bumpalo 3.1.1 (registry+https://github.com/rust-lang/crates.io-index)" = "8fe2567a8d8a3aedb4e39aa39e186d5673acfd56393c6ac83b2bc5bd82f4369c"
 "checksum bytes 0.5.3 (registry+https://github.com/rust-lang/crates.io-index)" = "10004c15deb332055f7a4a208190aed362cf9a7c2f6ab70a305fba50e1105f38"

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "monolith"
-version = "2.2.1"
+version = "2.2.4"
 edition = "2018"
 authors = [
     "Sunshine <sunshine@uberspace.net>",
@@ -12,7 +12,7 @@ authors = [
 description = "CLI tool for saving web pages as a single HTML file"
 
 [dependencies]
-base64 = "0.11.0"
+base64 = "0.12.0"
 clap = "2.33.0"
 cssparser = "0.27.2"
 html5ever = "0.24.1"

Makefile

@@ -1,21 +1,25 @@
-#!/usr/bin/make -f
+# Makefile for monolith
 
-all: test
+all: build
 .PHONY: all
 
 build:
 	@cargo build --locked
 .PHONY: build
 
-install:
-	@cargo install --force --locked --path .
-.PHONY: install
-
 test: build
 	@cargo test --locked
 	@cargo fmt --all -- --check
-.PHONY: test
+.PHONY: test_code_formatting
 
 lint:
 	@cargo fmt --all --
 .PHONY: lint
+
+install:
+	@cargo install --force --locked --path .
+.PHONY: install
+
+uninstall:
+	@cargo uninstall
+.PHONY: uninstall

README.md

@@ -1,5 +1,6 @@
-[![Travis CI Build Status](https://travis-ci.org/Y2Z/monolith.svg?branch=master)](https://travis-ci.org/Y2Z/monolith)
-[![AppVeyor Build status](https://ci.appveyor.com/api/projects/status/ae7soyjih8jg2bv7/branch/master?svg=true)](https://ci.appveyor.com/project/snshn/monolith/branch/master)
+[![Monolith Build Status for GNU/Linux](https://github.com/Y2Z/monolith/workflows/GNU%2FLinux/badge.svg)](https://github.com/Y2Z/monolith/actions?query=workflow%3AGNU%2FLinux)
+[![Monolith Build Status for macOS](https://github.com/Y2Z/monolith/workflows/macOS/badge.svg)](https://github.com/Y2Z/monolith/actions?query=workflow%3AmacOS)
+[![Monolith Build Status for Windows](https://github.com/Y2Z/monolith/workflows/Windows/badge.svg)](https://github.com/Y2Z/monolith/actions?query=workflow%3AWindows)
 
 ```
  ___     ___________    __________      ___________________    ___
@@ -11,9 +12,9 @@
 |___|   |__________|  \____________________|   |___|   |___|  |___|
 ```
 
-A data hoarder's dream come true: bundle any web page into a single HTML file. You can finally replace that gazillion of open tabs with a gazillion of .html files stored somewhere on your precious little drive.
+A data hoarder’s dream come true: bundle any web page into a single HTML file. You can finally replace that gazillion of open tabs with a gazillion of .html files stored somewhere on your precious little drive.
 
-Unlike the conventional "Save page as", `monolith` not only saves the target document, it embeds CSS, image, and JavaScript assets **all at once**, producing a single HTML5 document that is a joy to store and share.
+Unlike the conventional “Save page as”, `monolith` not only saves the target document, it embeds CSS, image, and JavaScript assets **all at once**, producing a single HTML5 document that is a joy to store and share.
 
 If compared to saving websites with `wget -mpk`, this tool embeds all assets as data URLs and therefore lets browsers render the saved page exactly the way it was on the Internet, even when no network connection is available.
 
@@ -21,18 +22,21 @@ If compared to saving websites with `wget -mpk`, this tool embeds all assets as
 
 ## Installation
 
-#### From source
-    $ git clone https://github.com/Y2Z/monolith.git
-    $ cd monolith
-    $ make install
-
-#### With Homebrew (on macOS and GNU/Linux)
+#### Via Homebrew (on macOS and GNU/Linux)
     $ brew install monolith
 
 #### Using Snapcraft (on GNU/Linux)
     $ snap install monolith
 
-#### Via Docker
+#### From source
+
+    Dependency: `libssl-dev`
+
+    $ git clone https://github.com/Y2Z/monolith.git
+    $ cd monolith
+    $ make install
+
+#### With Docker
 The guide can be found [here](docs/containers.md)
 
 ---------------------------------------------------
@@ -45,6 +49,7 @@ The guide can be found [here](docs/containers.md)
 ## Options
  - `-c`: Ignore styles
  - `-f`: Exclude frames and iframes
+ - `-F`: Omit web fonts
  - `-i`: Remove images
  - `-I`: Isolate the document
  - `-j`: Exclude JavaScript
@@ -79,4 +84,4 @@ The Unlicense
 ---------------------------------------------------
 
 <!-- Microtext -->
-<sub>Keep in mind that `monolith` is not aware of your browser's session</sub>
+<sub>Keep in mind that `monolith` is not aware of your browser’s session</sub>

docs/arch/0002-noscript-nodes.md

@@ -0,0 +1,19 @@
+# 2. NOSCRIPT nodes
+
+Date: 2020-04-16
+
+## Status
+
+Accepted
+
+## Context
+
+HTML pages sometimes contain NOSCRIPT nodes, which reveal their contents only in case when JavaScript is not available. Most of the time they contain hidden messages that inform about certain JavaScript-dependent features not being operational, however sometimes can also feature media assets or even iframes.
+
+## Decision
+
+When the document is being saved with or without JavaScript, each NOSCRIPT node should be preserved while its children need to be processed exactly the same way as the rest of the document. This approach will ensure that even hidden remote assets are embedded — since those hidden elements may have to be displayed later in a browser that has JavaScript turned off. An option should be available to "unwrap" all NOSCRIPT nodes in order to make their contents always visible in the document, complimenting the "disable JS" function of the program.
+
+## Consequences
+
+Saved documents will have contents of all NOSCRIPT nodes processed as if they are part of the document's DOM, therefore properly display images encapsulated within NOSCRIPT nodes when being viewed in browsers that have JavaScript turned off (or have no JavaScript support in the first place). The new option to "unwrap" NOSCRIPT elements will help the user ensure that the resulting document always represents what the original web page looked like in a browser that had JavaScript turned off.

docs/arch/0005-asset-minimization.md

@@ -21,5 +21,5 @@ saved by monolith, if needed.
 
 ## Consequences
 
-Monolith will not support modification of original document assets for the purpose of reducing their size, sticking to performing only a minimal
+Monolith will not support modification of original document assets for the purpose of reducing their size, sticking to performing only minimal
 amount of modifications to the original web page — whatever is needed to provide security or exclude unwanted asset types.

src/args.rs

@@ -2,8 +2,9 @@ use clap::{App, Arg};
 
 #[derive(Default)]
 pub struct AppArgs {
-    pub url_target: String,
+    pub target: String,
     pub no_css: bool,
+    pub no_fonts: bool,
     pub no_frames: bool,
     pub no_images: bool,
     pub no_js: bool,
@@ -26,15 +27,16 @@ impl AppArgs {
             .author(crate_authors!("\n"))
             .about(crate_description!())
             .arg(
-                Arg::with_name("url")
+                Arg::with_name("target")
                     .required(true)
                     .takes_value(true)
                     .index(1)
-                    .help("URL to download"),
+                    .help("URL or file path"),
             )
             // .args_from_usage("-a, --include-audio 'Removes audio sources'")
             .args_from_usage("-c, --no-css 'Removes CSS'")
             .args_from_usage("-f, --no-frames 'Removes frames and iframes'")
+            .args_from_usage("-F, --no-fonts 'Removes fonts'")
             .args_from_usage("-i, --no-images 'Removes images'")
             .args_from_usage("-I, --isolate 'Cuts off document from the Internet'")
             .args_from_usage("-j, --no-js 'Removes JavaScript'")
@@ -47,11 +49,12 @@ impl AppArgs {
             .get_matches();
         let mut app_args = AppArgs::default();
         // Process the command
-        app_args.url_target = app
-            .value_of("url")
-            .expect("please set target url")
+        app_args.target = app
+            .value_of("target")
+            .expect("please set target")
             .to_string();
         app_args.no_css = app.is_present("no-css");
+        app_args.no_fonts = app.is_present("no-fonts");
         app_args.no_frames = app.is_present("no-frames");
         app_args.no_images = app.is_present("no-images");
         app_args.no_js = app.is_present("no-js");

src/css.rs

@@ -2,12 +2,12 @@ use cssparser::{ParseError, Parser, ParserInput, SourcePosition, Token};
 use reqwest::blocking::Client;
 use std::collections::HashMap;
 
-use crate::utils::{data_to_data_url, decode_url, get_url_fragment, resolve_url, retrieve_asset};
+use crate::utils::{data_to_data_url, get_url_fragment, resolve_url, retrieve_asset};
 
 const CSS_PROPS_WITH_IMAGE_URLS: &[&str] = &[
+    // Universal
     "background",
     "background-image",
-    "border",
     "border-image",
     "border-image-source",
     "content",
@@ -16,10 +16,15 @@ const CSS_PROPS_WITH_IMAGE_URLS: &[&str] = &[
     "list-style-image",
     "mask",
     "mask-image",
+    // Specific to @counter-style
+    "additive-symbols",
+    "negative",
+    "pad",
+    "prefix",
+    "suffix",
+    "symbols",
 ];
-
-const TRANSPARENT_PIXEL: &str = "data:image/png;base64,\
-     iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=";
+const CSS_SPECIAL_CHARS: &str = "~!@$%^&*()+=,./'\";:?><[]{}|`#";
 
 pub fn is_image_url_prop(prop_name: &str) -> bool {
     CSS_PROPS_WITH_IMAGE_URLS
@@ -36,14 +41,27 @@ pub fn enquote(input: String, double: bool) -> String {
     }
 }
 
+pub fn escape(value: &str) -> String {
+    let mut res = str!(&value);
+
+    res = res.replace("\\", "\\\\");
+
+    for c in CSS_SPECIAL_CHARS.chars() {
+        res = res.replace(c, format!("\\{}", c).as_str());
+    }
+
+    res
+}
+
 pub fn process_css<'a>(
-    cache: &mut HashMap<String, String>,
+    cache: &mut HashMap<String, Vec<u8>>,
     client: &Client,
     parent_url: &str,
     parser: &mut Parser,
     rule_name: &str,
     prop_name: &str,
     func_name: &str,
+    opt_no_fonts: bool,
     opt_no_images: bool,
     opt_silent: bool,
 ) -> Result<String, ParseError<'a, String>> {
@@ -72,6 +90,10 @@ pub fn process_css<'a>(
             Token::Colon => result.push_str(":"),
             Token::Comma => result.push_str(","),
             Token::ParenthesisBlock | Token::SquareBracketBlock | Token::CurlyBracketBlock => {
+                if opt_no_fonts && curr_rule == "font-face" {
+                    continue;
+                }
+
                 let closure: &str;
                 if token == &Token::ParenthesisBlock {
                     result.push_str("(");
@@ -94,6 +116,7 @@ pub fn process_css<'a>(
                             rule_name,
                             curr_prop.as_str(),
                             func_name,
+                            opt_no_fonts,
                             opt_no_images,
                             opt_silent,
                         )
@@ -116,12 +139,18 @@ pub fn process_css<'a>(
             Token::WhiteSpace(ref value) => {
                 result.push_str(value);
             }
+            // div...
             Token::Ident(ref value) => {
+                curr_rule = str!();
                 curr_prop = str!(value);
-                result.push_str(value);
+                result.push_str(&escape(value));
             }
+            // @import, @font-face, @charset, @media...
             Token::AtKeyword(ref value) => {
                 curr_rule = str!(value);
+                if opt_no_fonts && curr_rule == "font-face" {
+                    continue;
+                }
                 result.push_str("@");
                 result.push_str(value);
             }
@@ -130,13 +159,10 @@ pub fn process_css<'a>(
                 result.push_str(value);
             }
             Token::QuotedString(ref value) => {
-                let is_import: bool = curr_rule == "import";
-                if is_import {
+                if curr_rule == "import" {
                     // Reset current at-rule value
                     curr_rule = str!();
-                }
 
-                if is_import {
                     // Skip empty import values
                     if value.len() < 1 {
                         result.push_str("''");
@@ -145,12 +171,11 @@ pub fn process_css<'a>(
 
                     let full_url = resolve_url(&parent_url, value).unwrap_or_default();
                     let url_fragment = get_url_fragment(full_url.clone());
-                    let full_url_decoded = decode_url(full_url);
                     let (css, final_url) = retrieve_asset(
                         cache,
                         client,
                         &parent_url,
-                        &full_url_decoded,
+                        &full_url,
                         false,
                         "",
                         opt_silent,
@@ -166,6 +191,7 @@ pub fn process_css<'a>(
                                     client,
                                     final_url.as_str(),
                                     &css,
+                                    opt_no_fonts,
                                     opt_no_images,
                                     opt_silent,
                                 )
@@ -185,7 +211,7 @@ pub fn process_css<'a>(
                         }
 
                         if opt_no_images && is_image_url_prop(curr_prop.as_str()) {
-                            result.push_str(enquote(str!(TRANSPARENT_PIXEL), false).as_str());
+                            result.push_str(enquote(str!(empty_image!()), false).as_str());
                         } else {
                             let resolved_url = resolve_url(&parent_url, value).unwrap_or_default();
                             let (data_url, _final_url) = retrieve_asset(
@@ -220,26 +246,33 @@ pub fn process_css<'a>(
                 ref unit_value,
                 ..
             } => {
-                if *has_sign {
-                    result.push_str("-");
+                if *has_sign && *unit_value >= 0. {
+                    result.push_str("+");
                 }
                 result.push_str(str!(unit_value * 100.).as_str());
                 result.push_str("%");
             }
             Token::Dimension {
+                ref has_sign,
                 ref value,
                 ref unit,
                 ..
             } => {
+                if *has_sign && *value >= 0. {
+                    result.push_str("+");
+                }
                 result.push_str(str!(value).as_str());
                 result.push_str(str!(unit).as_str());
             }
+            // #selector, #id...
             Token::IDHash(ref value) => {
+                curr_rule = str!();
                 result.push_str("#");
-                result.push_str(value);
+                result.push_str(&escape(value));
             }
             Token::UnquotedUrl(ref value) => {
                 let is_import: bool = curr_rule == "import";
+
                 if is_import {
                     // Reset current at-rule value
                     curr_rule = str!();
@@ -260,12 +293,11 @@ pub fn process_css<'a>(
                 if is_import {
                     let full_url = resolve_url(&parent_url, value).unwrap_or_default();
                     let url_fragment = get_url_fragment(full_url.clone());
-                    let full_url_decoded = decode_url(full_url);
                     let (css, final_url) = retrieve_asset(
                         cache,
                         client,
                         &parent_url,
-                        &full_url_decoded,
+                        &full_url,
                         false,
                         "",
                         opt_silent,
@@ -281,6 +313,7 @@ pub fn process_css<'a>(
                                     client,
                                     final_url.as_str(),
                                     &css,
+                                    opt_no_fonts,
                                     opt_no_images,
                                     opt_silent,
                                 )
@@ -294,7 +327,7 @@ pub fn process_css<'a>(
                     );
                 } else {
                     if opt_no_images && is_image_url_prop(curr_prop.as_str()) {
-                        result.push_str(enquote(str!(TRANSPARENT_PIXEL), false).as_str());
+                        result.push_str(enquote(str!(empty_image!()), false).as_str());
                     } else {
                         let full_url = resolve_url(&parent_url, value).unwrap_or_default();
                         let (data_url, _final_url) = retrieve_asset(
@@ -328,6 +361,7 @@ pub fn process_css<'a>(
                             curr_rule.as_str(),
                             curr_prop.as_str(),
                             function_name,
+                            opt_no_fonts,
                             opt_no_images,
                             opt_silent,
                         )
@@ -345,10 +379,11 @@ pub fn process_css<'a>(
 }
 
 pub fn embed_css(
-    cache: &mut HashMap<String, String>,
+    cache: &mut HashMap<String, Vec<u8>>,
     client: &Client,
     parent_url: &str,
     css: &str,
+    opt_no_fonts: bool,
     opt_no_images: bool,
     opt_silent: bool,
 ) -> String {
@@ -363,6 +398,7 @@ pub fn embed_css(
         "",
         "",
         "",
+        opt_no_fonts,
         opt_no_images,
         opt_silent,
     )

src/html.rs

@@ -20,38 +20,33 @@ const ICON_VALUES: &[&str] = &[
     "fluid-icon",
 ];
 
-const TRANSPARENT_PIXEL: &str = "data:image/png;base64,\
-     iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=";
-
 pub fn get_parent_node(node: &Handle) -> Handle {
     let parent = node.parent.take().clone();
     parent.and_then(|node| node.upgrade()).unwrap()
 }
 
-pub fn get_node_name(node: &Handle) -> &'_ str {
+pub fn get_node_name(node: &Handle) -> Option<&'_ str> {
     match &node.data {
-        NodeData::Element { ref name, .. } => name.local.as_ref(),
-        _ => "",
+        NodeData::Element { ref name, .. } => Some(name.local.as_ref()),
+        _ => None,
     }
 }
 
 pub fn is_icon(attr_value: &str) -> bool {
-    ICON_VALUES
-        .iter()
-        .find(|a| attr_value.eq_ignore_ascii_case(a))
-        .is_some()
+    ICON_VALUES.contains(&attr_value.to_lowercase().as_str())
 }
 
 pub fn walk_and_embed_assets(
-    cache: &mut HashMap<String, String>,
+    cache: &mut HashMap<String, Vec<u8>>,
     client: &Client,
     url: &str,
     node: &Handle,
     opt_no_css: bool,
+    opt_no_fonts: bool,
+    opt_no_frames: bool,
     opt_no_js: bool,
     opt_no_images: bool,
     opt_silent: bool,
-    opt_no_frames: bool,
 ) {
     match node.data {
         NodeData::Document => {
@@ -63,10 +58,11 @@ pub fn walk_and_embed_assets(
                     &url,
                     child,
                     opt_no_css,
+                    opt_no_fonts,
+                    opt_no_frames,
                     opt_no_js,
                     opt_no_images,
                     opt_silent,
-                    opt_no_frames,
                 );
             }
         }
@@ -163,17 +159,18 @@ pub fn walk_and_embed_assets(
                                         ) {
                                             // On successful retrieval, traverse CSS
                                             Ok((css_data, final_url)) => {
-                                                let x: String = embed_css(
+                                                let css: String = embed_css(
                                                     cache,
                                                     client,
                                                     &final_url,
                                                     &css_data,
+                                                    opt_no_fonts,
                                                     opt_no_images,
                                                     opt_silent,
                                                 );
                                                 data_to_data_url(
                                                     "text/css",
-                                                    x.as_bytes(),
+                                                    css.as_bytes(),
                                                     &final_url,
                                                     "",
                                                 )
@@ -273,7 +270,7 @@ pub fn walk_and_embed_assets(
                     if opt_no_images {
                         attrs_mut.push(Attribute {
                             name: QualName::new(None, ns!(), local_name!("src")),
-                            value: Tendril::from_slice(TRANSPARENT_PIXEL),
+                            value: Tendril::from_slice(empty_image!()),
                         });
                     } else if let Some((data_url, _)) = found_datasrc
                         .iter()
@@ -300,6 +297,63 @@ pub fn walk_and_embed_assets(
                         });
                     }
                 }
+                "input" => {
+                    let mut is_image: bool = false;
+                    for attr in attrs_mut.iter_mut() {
+                        let attr_name: &str = &attr.name.local;
+                        if attr_name == "type" {
+                            is_image = attr.value.to_string().eq_ignore_ascii_case("image");
+                        }
+                    }
+
+                    if is_image {
+                        let mut found_src: Option<Attribute> = None;
+                        let mut i = 0;
+                        while i < attrs_mut.len() {
+                            let attr_name = attrs_mut[i].name.local.as_ref();
+                            if attr_name.eq_ignore_ascii_case("src") {
+                                found_src = Some(attrs_mut.remove(i));
+                            } else {
+                                i += 1;
+                            }
+                        }
+
+                        // If images are disabled, clear both sources
+                        if opt_no_images {
+                            attrs_mut.push(Attribute {
+                                name: QualName::new(None, ns!(), local_name!("src")),
+                                value: Tendril::from_slice(empty_image!()),
+                            });
+                        } else if let Some((data_url, _)) = found_src
+                            .iter()
+                            .map(|attr| attr.value.trim())
+                            .filter(|src| !src.is_empty()) // Skip if empty
+                            .next()
+                            .and_then(|src| resolve_url(&url, src).ok()) // Make absolute
+                            .and_then(|abs_src| // Download and convert to data_url
+                                retrieve_asset(
+                                    cache,
+                                    client,
+                                    &url,
+                                    &abs_src,
+                                    true,
+                                    "",
+                                    opt_silent,
+                                ).ok())
+                        {
+                            // Add new data_url src attribute
+                            attrs_mut.push(Attribute {
+                                name: QualName::new(None, ns!(), local_name!("src")),
+                                value: Tendril::from_slice(data_url.as_ref()),
+                            });
+                        }
+                    }
+                }
+                "svg" => {
+                    if opt_no_images {
+                        node.children.borrow_mut().clear();
+                    }
+                }
                 "source" => {
                     for attr in attrs_mut.iter_mut() {
                         let attr_name: &str = &attr.name.local;
@@ -310,10 +364,10 @@ pub fn walk_and_embed_assets(
                             attr.value.clear();
                             attr.value.push_slice(src_full_url.as_str());
                         } else if attr_name == "srcset" {
-                            if get_node_name(&get_parent_node(&node)) == "picture" {
+                            if get_node_name(&get_parent_node(&node)) == Some("picture") {
                                 if opt_no_images {
                                     attr.value.clear();
-                                    attr.value.push_slice(TRANSPARENT_PIXEL);
+                                    attr.value.push_slice(empty_image!());
                                 } else {
                                     let srcset_full_url =
                                         resolve_url(&url, attr.value.trim()).unwrap_or_default();
@@ -334,7 +388,7 @@ pub fn walk_and_embed_assets(
                         }
                     }
                 }
-                "a" => {
+                "a" | "area" => {
                     for attr in attrs_mut.iter_mut() {
                         if &attr.name.local == "href" {
                             let attr_value = attr.value.trim();
@@ -411,6 +465,7 @@ pub fn walk_and_embed_assets(
                                     client,
                                     &url,
                                     tendril.as_ref(),
+                                    opt_no_fonts,
                                     opt_no_images,
                                     opt_silent,
                                 );
@@ -468,10 +523,11 @@ pub fn walk_and_embed_assets(
                                 &frame_final_url,
                                 &dom.document,
                                 opt_no_css,
+                                opt_no_fonts,
+                                opt_no_frames,
                                 opt_no_js,
                                 opt_no_images,
                                 opt_silent,
-                                opt_no_frames,
                             );
                             let mut buf: Vec<u8> = Vec::new();
                             serialize(&mut buf, &dom.document, SerializeOpts::default()).unwrap();
@@ -539,6 +595,7 @@ pub fn walk_and_embed_assets(
                         client,
                         &url,
                         attribute.value.as_ref(),
+                        opt_no_fonts,
                         opt_no_images,
                         opt_silent,
                     );
@@ -570,10 +627,11 @@ pub fn walk_and_embed_assets(
                     &url,
                     child,
                     opt_no_css,
+                    opt_no_fonts,
+                    opt_no_frames,
                     opt_no_js,
                     opt_no_images,
                     opt_silent,
-                    opt_no_frames,
                 );
             }
         }

src/macros.rs

@@ -7,3 +7,11 @@ macro_rules! str {
         ToString::to_string(&$val)
     };
 }
+
+#[macro_export]
+macro_rules! empty_image {
+    () => {
+"data:image/png;base64,\
+iVBORw0KGgoAAAANSUhEUgAAAA0AAAANCAQAAADY4iz3AAAAEUlEQVR42mNkwAkYR6UolgIACvgADsuK6xYAAAAASUVORK5CYII="
+    };
+}

src/main.rs

@@ -47,7 +47,7 @@ impl Output {
 
 fn main() {
     let app_args = AppArgs::get();
-    let original_target: &str = &app_args.url_target;
+    let original_target: &str = &app_args.target;
     let target_url: &str;
     let base_url;
     let dom;
@@ -122,9 +122,9 @@ fn main() {
         base_url = final_url;
         dom = html_to_dom(&data);
     } else if is_data_url(target_url) {
-        let text: String = data_url_to_text(target_url);
-        if text.len() == 0 {
-            eprintln!("Unsupported data URL input");
+        let (media_type, text): (String, String) = data_url_to_text(target_url);
+        if !media_type.eq_ignore_ascii_case("text/html") {
+            eprintln!("Unsupported data URL media type");
             process::exit(1);
         }
         base_url = str!(target_url);
@@ -139,10 +139,11 @@ fn main() {
         &base_url,
         &dom.document,
         app_args.no_css,
+        app_args.no_fonts,
+        app_args.no_frames,
         app_args.no_js,
         app_args.no_images,
         app_args.silent,
-        app_args.no_frames,
     );
 
     let html: String = stringify_document(

src/tests/cli.rs

@@ -62,7 +62,7 @@ fn passing_bad_input_data_url() -> Result<(), Box<dyn std::error::Error>> {
     // STDERR should contain error description
     assert_eq!(
         std::str::from_utf8(&out.stderr).unwrap(),
-        "Unsupported data URL input\n"
+        "Unsupported data URL media type\n"
     );
 
     // The exit code should be 1
@@ -162,15 +162,18 @@ fn passing_remove_images_from_data_url() -> Result<(), Box<dyn std::error::Error
     // STDOUT should contain HTML with no images
     assert_eq!(
         std::str::from_utf8(&out.stdout).unwrap(),
-        "<html>\
+        format!(
+            "<html>\
 <head>\
 <meta http-equiv=\"Content-Security-Policy\" content=\"img-src data:;\"></meta>\
 </head>\
 <body>\
-<img src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=\">\
+<img src=\"{empty_image}\">\
 Hi\
 </body>\
-</html>\n"
+</html>\n",
+            empty_image = empty_image!()
+        )
     );
 
     // STDERR should be empty
@@ -229,7 +232,8 @@ fn passing_local_file_target_input() -> Result<(), Box<dyn std::error::Error>> {
     // STDOUT should contain HTML from the local file
     assert_eq!(
         std::str::from_utf8(&out.stdout).unwrap(),
-        "<!DOCTYPE html><html lang=\"en\"><head>\n  \
+        "\
+<!DOCTYPE html><html lang=\"en\"><head>\n  \
 <meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">\n  \
 <title>Local HTML file</title>\n  \
 <link href=\"data:text/css;base64,Ym9keSB7CiAgICBiYWNrZ3JvdW5kLWNvbG9yOiAjMDAwOwogICAgY29sb3I6ICNmZmY7Cn0K\" rel=\"stylesheet\" type=\"text/css\">\n  \
@@ -238,16 +242,19 @@ fn passing_local_file_target_input() -> Result<(), Box<dyn std::error::Error>> {
 <a href=\"file://local-file.html/\">Tricky href</a>\n  \
 <a href=\"https://github.com/Y2Z/monolith\">Remote URL</a>\n  \
 <script src=\"data:application/javascript;base64,ZG9jdW1lbnQuYm9keS5zdHlsZS5iYWNrZ3JvdW5kQ29sb3IgPSAiZ3JlZW4iOwpkb2N1bWVudC5ib2R5LnN0eWxlLmNvbG9yID0gInJlZCI7Cg==\"></script>\n\n\n\n\
-</body></html>\n"
+</body></html>\n\
+"
     );
 
     // STDERR should contain list of retrieved file URLs
     assert_eq!(
         std::str::from_utf8(&out.stderr).unwrap(),
         format!(
-            "{file}{cwd}/src/tests/data/local-file.html\n\
+            "\
+{file}{cwd}/src/tests/data/local-file.html\n\
 {file}{cwd}/src/tests/data/local-style.css\n\
-{file}{cwd}/src/tests/data/local-script.js\n",
+{file}{cwd}/src/tests/data/local-script.js\n\
+",
             file = file_url_protocol,
             cwd = cwd_normalized
         )
@@ -286,17 +293,22 @@ fn passing_local_file_target_input_absolute_target_path() -> Result<(), Box<dyn
     // STDOUT should contain HTML from the local file
     assert_eq!(
         std::str::from_utf8(&out.stdout).unwrap(),
-        "<!DOCTYPE html><html lang=\"en\"><head>\
+        format!(
+            "\
+<!DOCTYPE html><html lang=\"en\"><head>\
 <meta http-equiv=\"Content-Security-Policy\" content=\"default-src 'unsafe-inline' data:; style-src 'none'; script-src 'none'; img-src data:;\"></meta>\n  \
 <meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">\n  \
 <title>Local HTML file</title>\n  \
 <link href=\"\" rel=\"stylesheet\" type=\"text/css\">\n  \
 <link href=\"\" rel=\"stylesheet\" type=\"text/css\">\n</head>\n\n<body>\n  \
-<img alt=\"\" src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=\">\n  \
+<img alt=\"\" src=\"{empty_image}\">\n  \
 <a href=\"file://local-file.html/\">Tricky href</a>\n  \
 <a href=\"https://github.com/Y2Z/monolith\">Remote URL</a>\n  \
 <script src=\"\"></script>\n\n\n\n\
-</body></html>\n"
+</body></html>\n\
+",
+            empty_image = empty_image!()
+        )
     );
 
     // STDERR should contain only the target file
@@ -342,17 +354,22 @@ fn passing_local_file_url_target_input() -> Result<(), Box<dyn std::error::Error
     // STDOUT should contain HTML from the local file
     assert_eq!(
         std::str::from_utf8(&out.stdout).unwrap(),
-        "<!DOCTYPE html><html lang=\"en\"><head>\
+        format!(
+            "\
+<!DOCTYPE html><html lang=\"en\"><head>\
 <meta http-equiv=\"Content-Security-Policy\" content=\"style-src 'none'; script-src 'none'; img-src data:;\"></meta>\n  \
 <meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">\n  \
 <title>Local HTML file</title>\n  \
 <link href=\"\" rel=\"stylesheet\" type=\"text/css\">\n  \
 <link href=\"\" rel=\"stylesheet\" type=\"text/css\">\n</head>\n\n<body>\n  \
-<img alt=\"\" src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=\">\n  \
+<img alt=\"\" src=\"{empty_image}\">\n  \
 <a href=\"file://local-file.html/\">Tricky href</a>\n  \
 <a href=\"https://github.com/Y2Z/monolith\">Remote URL</a>\n  \
 <script src=\"\"></script>\n\n\n\n\
-</body></html>\n"
+</body></html>\n\
+",
+            empty_image = empty_image!()
+        )
     );
 
     // STDERR should contain list of retrieved file URLs
@@ -458,7 +475,8 @@ fn passing_css_import_string() -> Result<(), Box<dyn std::error::Error>> {
     let mut file_html = NamedTempFile::new()?;
     writeln!(
         file_html,
-        "<style>\n\
+        "\
+<style>\n\
 @charset 'UTF-8';\n\
 \n\
 @import '{file}{css_path}';\n\
@@ -466,7 +484,8 @@ fn passing_css_import_string() -> Result<(), Box<dyn std::error::Error>> {
 @import url({file}{css_path});\n\
 \n\
 @import url('{file}{css_path}')\n\
-</style>\n",
+</style>\n\
+",
         file = file_url_prefix,
         css_path = str!(file_css.path().to_str().unwrap()).replace("\\", "/"),
     )?;

src/tests/css/embed_css.rs

@@ -15,7 +15,10 @@ fn passing_empty_input() {
     let cache = &mut HashMap::new();
     let client = Client::new();
 
-    assert_eq!(css::embed_css(cache, &client, "", "", false, false,), "");
+    assert_eq!(
+        css::embed_css(cache, &client, "", "", false, false, false,),
+        ""
+    );
 }
 
 #[test]
@@ -37,16 +40,20 @@ height: calc(100vh - 10pt)";
             &client,
             "https://doesntmatter.local/",
             &STYLE,
+            false,
             true,
             true,
         ),
-        "/* border: none;*/\
-background-image: url('data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII='); \
-list-style: url('data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=');\
+        format!(
+            "/* border: none;*/\
+background-image: url('{empty_image}'); \
+list-style: url('{empty_image}');\
 width:99.998%; \
 margin-top: -20px; \
 line-height: -1; \
-height: calc(100vh - 10pt)"
+height: calc(100vh - 10pt)",
+            empty_image = empty_image!()
+        )
     );
 }
 
@@ -64,21 +71,17 @@ line-height: -1; \
 height: calc(100vh - 10pt)";
 
     assert_eq!(
-        css::embed_css(
-            cache,
-            &client,
-            "",
-            &STYLE,
-            true,
-            true,
-        ),
-        "/* border: none;*/\
-background-image: url('data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII='); \
-list-style: url('data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=');\
+        css::embed_css(cache, &client, "", &STYLE, false, true, true,),
+        format!(
+            "/* border: none;*/\
+background-image: url('{empty_image}'); \
+list-style: url('{empty_image}');\
 width:99.998%; \
 margin-top: -20px; \
 line-height: -1; \
-height: calc(100vh - 10pt)"
+height: calc(100vh - 10pt)",
+            empty_image = empty_image!()
+        )
     );
 }
 
@@ -96,7 +99,7 @@ fn passing_style_block() {
 html > body {}";
 
     assert_eq!(
-        css::embed_css(cache, &client, "file:///", &CSS, false, true,),
+        css::embed_css(cache, &client, "file:///", &CSS, false, false, true,),
         CSS
     );
 }
@@ -136,7 +139,10 @@ fn passing_attribute_selectors() {
 }
 ";
 
-    assert_eq!(css::embed_css(cache, &client, "", &CSS, false, false,), CSS);
+    assert_eq!(
+        css::embed_css(cache, &client, "", &CSS, false, false, false,),
+        CSS
+    );
 }
 
 #[test]
@@ -159,14 +165,15 @@ fn passing_import_string() {
             "https://doesntmatter.local/",
             &CSS,
             false,
+            false,
             true,
         ),
         "\
 @charset 'UTF-8';\n\
 \n\
-@import 'data:text/css;base64,ZGF0YTp0ZXh0L2NzcyxodG1se2JhY2tncm91bmQtY29sb3I6IzAwMH0=';\n\
+@import 'data:text/css;base64,aHRtbHtiYWNrZ3JvdW5kLWNvbG9yOiMwMDB9';\n\
 \n\
-@import url('data:text/css;base64,ZGF0YTp0ZXh0L2NzcyxodG1se2NvbG9yOiNmZmZ9')\n\
+@import url('data:text/css;base64,aHRtbHtjb2xvcjojZmZmfQ==')\n\
 "
     );
 }
@@ -193,8 +200,118 @@ body {\n    \
             "https://doesntmatter.local/",
             &CSS,
             false,
+            false,
             true,
         ),
         CSS
     );
 }
+
+#[test]
+fn passing_transform_percentages_and_degrees() {
+    let cache = &mut HashMap::new();
+    let client = Client::new();
+
+    const CSS: &str = "\
+div {\n    \
+    transform: translate(-50%, -50%) rotate(-45deg);\n\
+    transform: translate(50%, 50%) rotate(45deg);\n\
+    transform: translate(+50%, +50%) rotate(+45deg);\n\
+}\n\
+";
+
+    assert_eq!(
+        css::embed_css(
+            cache,
+            &client,
+            "https://doesntmatter.local/",
+            &CSS,
+            false,
+            false,
+            true,
+        ),
+        CSS
+    );
+}
+
+#[test]
+fn passing_unusual_indents() {
+    let cache = &mut HashMap::new();
+    let client = Client::new();
+
+    const CSS: &str = "\
+.is\\:good:hover {\n    \
+    color: green\n\
+}\n\
+\n\
+#\\~\\!\\@\\$\\%\\^\\&\\*\\(\\)\\+\\=\\,\\.\\/\\\\\\'\\\"\\;\\:\\?\\>\\<\\[\\]\\{\\}\\|\\`\\# {\n    \
+    color: black\n\
+}\n\
+";
+
+    assert_eq!(
+        css::embed_css(
+            cache,
+            &client,
+            "https://doesntmatter.local/",
+            &CSS,
+            false,
+            false,
+            true,
+        ),
+        CSS
+    );
+}
+
+#[test]
+fn passing_exclude_fonts() {
+    let cache = &mut HashMap::new();
+    let client = Client::new();
+
+    const CSS: &str = "\
+@font-face {\n    \
+    font-family: 'My Font';\n    \
+    src: url(my_font.woff);\n\
+}\n\
+\n\
+#identifier {\n    \
+    font-family: 'My Font' Arial\n\
+}\n\
+\n\
+@font-face {\n    \
+    font-family: 'My Font';\n    \
+    src: url(my_font.woff);\n\
+}\n\
+\n\
+div {\n    \
+    font-family: 'My Font' Verdana\n\
+}\n\
+";
+
+    const CSS_OUT: &str = " \
+\n\
+\n\
+#identifier {\n    \
+    font-family: 'My Font' Arial\n\
+}\n\
+\n \
+\n\
+\n\
+div {\n    \
+    font-family: 'My Font' Verdana\n\
+}\n\
+";
+
+    assert_eq!(
+        css::embed_css(
+            cache,
+            &client,
+            "https://doesntmatter.local/",
+            &CSS,
+            true,
+            false,
+            true,
+        ),
+        CSS_OUT
+    );
+}

src/tests/html/get_node_name.rs

@@ -28,11 +28,11 @@ fn get_node_name() {
                 let parent = html::get_parent_node(node);
                 let parent_node_name = html::get_node_name(&parent);
                 if node_name == "head" || node_name == "body" {
-                    assert_eq!(parent_node_name, "html");
+                    assert_eq!(parent_node_name, Some("html"));
                 } else if node_name == "div" {
-                    assert_eq!(parent_node_name, "body");
+                    assert_eq!(parent_node_name, Some("body"));
                 } else if node_name == "p" {
-                    assert_eq!(parent_node_name, "div");
+                    assert_eq!(parent_node_name, Some("div"));
                 }
 
                 for child in node.children.borrow().iter() {

src/tests/html/walk_and_embed_assets.rs

@@ -19,6 +19,7 @@ fn passing_basic() {
     let url = "http://localhost";
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = false;
     let opt_no_images: bool = false;
@@ -32,10 +33,11 @@ fn passing_basic() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -55,6 +57,7 @@ fn passing_ensure_no_recursive_iframe() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = false;
     let opt_no_images: bool = false;
@@ -68,10 +71,11 @@ fn passing_ensure_no_recursive_iframe() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -91,6 +95,7 @@ fn passing_ensure_no_recursive_frame() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = false;
     let opt_no_images: bool = false;
@@ -104,10 +109,11 @@ fn passing_ensure_no_recursive_frame() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -129,6 +135,7 @@ fn passing_no_css() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = true;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = false;
     let opt_no_images: bool = false;
@@ -141,10 +148,11 @@ fn passing_no_css() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -173,6 +181,7 @@ fn passing_no_images() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = false;
     let opt_no_images: bool = true;
@@ -186,10 +195,11 @@ fn passing_no_images() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -197,18 +207,19 @@ fn passing_no_images() {
 
     assert_eq!(
         buf.iter().map(|&c| c as char).collect::<String>(),
-        "<html>\
+        format!(
+            "<html>\
          <head>\
          <link rel=\"icon\" href=\"\">\
          </head>\
          <body>\
          <div>\
-         <img src=\"data:image/png;base64,\
-         iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0\
-         lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=\">\
+         <img src=\"{empty_image}\">\
          </div>\
          </body>\
-         </html>"
+         </html>",
+            empty_image = empty_image!()
+        )
     );
 }
 
@@ -220,6 +231,7 @@ fn passing_no_body_background_images() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = false;
     let opt_no_images: bool = true;
@@ -233,10 +245,11 @@ fn passing_no_body_background_images() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -256,6 +269,7 @@ fn passing_no_frames() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = true;
     let opt_no_js: bool = false;
     let opt_no_images: bool = false;
@@ -268,10 +282,11 @@ fn passing_no_frames() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -291,6 +306,7 @@ fn passing_no_iframes() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = true;
     let opt_no_js: bool = false;
     let opt_no_images: bool = false;
@@ -303,10 +319,11 @@ fn passing_no_iframes() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -329,6 +346,7 @@ fn passing_no_js() {
     let cache = &mut HashMap::new();
 
     let opt_no_css: bool = false;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = false;
     let opt_no_js: bool = true;
     let opt_no_images: bool = false;
@@ -342,10 +360,11 @@ fn passing_no_js() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();
@@ -368,6 +387,7 @@ fn passing_with_no_integrity() {
     let cache = &mut HashMap::new();
     let client = Client::new();
     let opt_no_css: bool = true;
+    let opt_no_fonts: bool = false;
     let opt_no_frames: bool = true;
     let opt_no_js: bool = true;
     let opt_no_images: bool = true;
@@ -379,10 +399,11 @@ fn passing_with_no_integrity() {
         &url,
         &dom.document,
         opt_no_css,
+        opt_no_fonts,
+        opt_no_frames,
         opt_no_js,
         opt_no_images,
         opt_silent,
-        opt_no_frames,
     );
 
     let mut buf: Vec<u8> = Vec::new();

src/tests/macros/empty_image.rs

@@ -0,0 +1,14 @@
+//  ██████╗  █████╗ ███████╗███████╗██╗███╗   ██╗ ██████╗
+//  ██╔══██╗██╔══██╗██╔════╝██╔════╝██║████╗  ██║██╔════╝
+//  ██████╔╝███████║███████╗███████╗██║██╔██╗ ██║██║  ███╗
+//  ██╔═══╝ ██╔══██║╚════██║╚════██║██║██║╚██╗██║██║   ██║
+//  ██║     ██║  ██║███████║███████║██║██║ ╚████║╚██████╔╝
+//  ╚═╝     ╚═╝  ╚═╝╚══════╝╚══════╝╚═╝╚═╝  ╚═══╝ ╚═════╝
+
+#[cfg(test)]
+mod passing {
+    #[test]
+    fn contains_correct_image_data() {
+        assert_eq!(empty_image!(), "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAA0AAAANCAQAAADY4iz3AAAAEUlEQVR42mNkwAkYR6UolgIACvgADsuK6xYAAAAASUVORK5CYII=");
+    }
+}

src/tests/macros/mod.rs

@@ -0,0 +1,2 @@
+mod empty_image;
+mod str;

src/tests/macros/str.rs

@@ -0,0 +1,24 @@
+//  ██████╗  █████╗ ███████╗███████╗██╗███╗   ██╗ ██████╗
+//  ██╔══██╗██╔══██╗██╔════╝██╔════╝██║████╗  ██║██╔════╝
+//  ██████╔╝███████║███████╗███████╗██║██╔██╗ ██║██║  ███╗
+//  ██╔═══╝ ██╔══██║╚════██║╚════██║██║██║╚██╗██║██║   ██║
+//  ██║     ██║  ██║███████║███████║██║██║ ╚████║╚██████╔╝
+//  ╚═╝     ╚═╝  ╚═╝╚══════╝╚══════╝╚═╝╚═╝  ╚═══╝ ╚═════╝
+
+#[cfg(test)]
+mod passing {
+    #[test]
+    fn returns_empty_string() {
+        assert_eq!(str!(), "");
+    }
+
+    #[test]
+    fn converts_integer_into_string() {
+        assert_eq!(str!(123), "123");
+    }
+
+    #[test]
+    fn converts_str_into_string() {
+        assert_eq!(str!("abc"), "abc");
+    }
+}

src/tests/mod.rs

@@ -2,4 +2,5 @@ mod cli;
 mod css;
 mod html;
 mod js;
+mod macros;
 mod utils;

src/tests/utils/clean_url.rs

@@ -30,3 +30,11 @@ fn passing_removes_empty_query_and_empty_fragment() {
         "https://somewhere.com/font.eot"
     );
 }
+
+#[test]
+fn passing_removes_empty_query_amp_and_empty_fragment() {
+    assert_eq!(
+        utils::clean_url("https://somewhere.com/font.eot?a=b&#"),
+        "https://somewhere.com/font.eot?a=b"
+    );
+}

src/tests/utils/data_url_to_text.rs

@@ -9,42 +9,76 @@ use crate::utils;
 
 #[test]
 fn passing_parse_text_html_base64() {
+    let (media_type, text) = utils::data_url_to_text("data:text/html;base64,V29yayBleHBhbmRzIHNvIGFzIHRvIGZpbGwgdGhlIHRpbWUgYXZhaWxhYmxlIGZvciBpdHMgY29tcGxldGlvbg==");
+
+    assert_eq!(media_type, "text/html");
     assert_eq!(
-        utils::data_url_to_text("data:text/html;base64,V29yayBleHBhbmRzIHNvIGFzIHRvIGZpbGwgdGhlIHRpbWUgYXZhaWxhYmxlIGZvciBpdHMgY29tcGxldGlvbg=="),
+        text,
         "Work expands so as to fill the time available for its completion"
     );
 }
 
 #[test]
 fn passing_parse_text_html_utf8() {
+    let (media_type, text) = utils::data_url_to_text(
+        "data:text/html;utf8,Work expands so as to fill the time available for its completion",
+    );
+
+    assert_eq!(media_type, "text/html");
     assert_eq!(
-        utils::data_url_to_text(
-            "data:text/html;utf8,Work expands so as to fill the time available for its completion"
-        ),
+        text,
         "Work expands so as to fill the time available for its completion"
     );
 }
 
 #[test]
 fn passing_parse_text_html_plaintext() {
+    let (media_type, text) = utils::data_url_to_text(
+        "data:text/html,Work expands so as to fill the time available for its completion",
+    );
+
+    assert_eq!(media_type, "text/html");
     assert_eq!(
-        utils::data_url_to_text(
-            "data:text/html,Work expands so as to fill the time available for its completion"
-        ),
+        text,
         "Work expands so as to fill the time available for its completion"
     );
 }
 
 #[test]
 fn passing_parse_text_html_charset_utf_8_between_two_whitespaces() {
+    let (media_type, text) = utils::data_url_to_text(" data:text/html;charset=utf-8,Work expands so as to fill the time available for its completion ");
+
+    assert_eq!(media_type, "text/html");
     assert_eq!(
-        utils::data_url_to_text(
-            " data:text/html;charset=utf-8,Work expands so as to fill the time available for its completion "
-        ),
+        text,
         "Work expands so as to fill the time available for its completion"
     );
 }
 
+#[test]
+fn passing_parse_text_css_url_encoded() {
+    let (media_type, text) = utils::data_url_to_text("data:text/css,div{background-color:%23000}");
+
+    assert_eq!(media_type, "text/css");
+    assert_eq!(text, "div{background-color:#000}");
+}
+
+#[test]
+fn passing_parse_no_media_type_base64() {
+    let (media_type, text) = utils::data_url_to_text("data:;base64,dGVzdA==");
+
+    assert_eq!(media_type, "");
+    assert_eq!(text, "test");
+}
+
+#[test]
+fn passing_parse_no_media_type_no_encoding() {
+    let (media_type, text) = utils::data_url_to_text("data:;,test%20test");
+
+    assert_eq!(media_type, "");
+    assert_eq!(text, "test test");
+}
+
 //  ███████╗ █████╗ ██╗██╗     ██╗███╗   ██╗ ██████╗
 //  ██╔════╝██╔══██╗██║██║     ██║████╗  ██║██╔════╝
 //  █████╗  ███████║██║██║     ██║██╔██╗ ██║██║  ███╗
@@ -54,5 +88,8 @@ fn passing_parse_text_html_charset_utf_8_between_two_whitespaces() {
 
 #[test]
 fn failing_just_word_data() {
-    assert_eq!(utils::data_url_to_text("data"), "");
+    let (media_type, text) = utils::data_url_to_text("data");
+
+    assert_eq!(media_type, "");
+    assert_eq!(text, "");
 }

src/tests/utils/decode_url.rs

@@ -24,3 +24,13 @@ fn passing_decode_file_url() {
         "file:///tmp/space here/test#1.html"
     );
 }
+
+#[test]
+fn passing_plus_sign() {
+    assert_eq!(
+        utils::decode_url(str!(
+            "fonts.somewhere.com/css?family=Open+Sans:300,400,400italic,600,600italic"
+        )),
+        "fonts.somewhere.com/css?family=Open+Sans:300,400,400italic,600,600italic"
+    );
+}

src/tests/utils/file_url_to_fs_path.rs

@@ -21,3 +21,18 @@ fn passing_remove_protocl_and_fragment() {
         );
     }
 }
+
+#[test]
+fn passing_decodes_urls() {
+    if cfg!(windows) {
+        assert_eq!(
+            utils::file_url_to_fs_path("file:///C:/Documents%20and%20Settings/some-file.html"),
+            "C:\\Documents and Settings\\some-file.html"
+        );
+    } else {
+        assert_eq!(
+            utils::file_url_to_fs_path("file:///home/user/My%20Documents"),
+            "/home/user/My Documents"
+        );
+    }
+}

src/utils.rs

@@ -30,6 +30,14 @@ const MAGIC: [[&[u8]; 2]; 18] = [
     [b"\x1A\x45\xDF\xA3", b"video/webm"],
 ];
 
+const PLAINTEXT_MEDIA_TYPES: &[&str] = &[
+    "image/svg+xml",
+    "text/css",
+    "text/html",
+    "text/javascript",
+    "text/plain",
+];
+
 pub fn data_to_data_url(media_type: &str, data: &[u8], url: &str, fragment: &str) -> String {
     let media_type: String = if media_type.is_empty() {
         detect_media_type(data, &url)
@@ -88,6 +96,10 @@ pub fn is_http_url<T: AsRef<str>>(url: T) -> bool {
         .unwrap_or(false)
 }
 
+pub fn is_plaintext_media_type(media_type: &str) -> bool {
+    PLAINTEXT_MEDIA_TYPES.contains(&media_type.to_lowercase().as_str())
+}
+
 pub fn resolve_url<T: AsRef<str>, U: AsRef<str>>(from: T, to: U) -> Result<String, ParseError> {
     let result = if is_http_url(to.as_ref()) {
         to.as_ref().to_string()
@@ -108,65 +120,72 @@ pub fn get_url_fragment<T: AsRef<str>>(url: T) -> String {
     }
 }
 
-pub fn clean_url<T: AsRef<str>>(url: T) -> String {
-    let mut result = Url::parse(url.as_ref()).unwrap();
+pub fn clean_url<T: AsRef<str>>(input: T) -> String {
+    let mut url = Url::parse(input.as_ref()).unwrap();
 
     // Clear fragment
-    result.set_fragment(None);
+    url.set_fragment(None);
 
     // Get rid of stray question mark
-    if result.query() == Some("") {
-        result.set_query(None);
+    if url.query() == Some("") {
+        url.set_query(None);
     }
-    result.to_string()
+
+    // Remove empty trailing ampersand(s)
+    let mut result: String = url.to_string();
+    while result.ends_with("&") {
+        result.pop();
+    }
+
+    result
 }
 
-pub fn data_url_to_text<T: AsRef<str>>(url: T) -> String {
-    let parsed_url = Url::parse(url.as_ref()).unwrap_or(Url::parse("http://[::1]").unwrap());
+pub fn data_url_to_text<T: AsRef<str>>(url: T) -> (String, String) {
+    let parsed_url = Url::parse(url.as_ref()).unwrap_or(Url::parse("data:,").unwrap());
     let path: String = parsed_url.path().to_string();
     let comma_loc: usize = path.find(',').unwrap_or(path.len());
 
-    if comma_loc == path.len() {
-        return str!();
-    }
-
     let meta_data: String = path.chars().take(comma_loc).collect();
     let raw_data: String = path.chars().skip(comma_loc + 1).collect();
 
     let data: String = decode_url(raw_data);
 
     let meta_data_items: Vec<&str> = meta_data.split(';').collect();
-    let mut media_type: &str = "";
     let mut encoding: &str = "";
 
+    let mut media_type: String = str!();
+    let mut text: String = str!();
+
     let mut i: i8 = 0;
     for item in &meta_data_items {
         if i == 0 {
-            if item.eq_ignore_ascii_case("text/html") {
-                media_type = item;
-                continue;
+            media_type = str!(item);
+        } else {
+            if item.eq_ignore_ascii_case("base64")
+                || item.eq_ignore_ascii_case("utf8")
+                || item.eq_ignore_ascii_case("charset=UTF-8")
+            {
+                encoding = item;
             }
         }
 
-        if item.eq_ignore_ascii_case("base64") || item.eq_ignore_ascii_case("utf8") {
-            encoding = item;
-        }
-
         i = i + 1;
     }
 
-    if media_type.eq_ignore_ascii_case("text/html") {
+    if is_plaintext_media_type(&media_type) || media_type.is_empty() {
         if encoding.eq_ignore_ascii_case("base64") {
-            String::from_utf8(base64::decode(&data).unwrap_or(vec![])).unwrap_or(str!())
+            text = String::from_utf8(base64::decode(&data).unwrap_or(vec![])).unwrap_or(str!())
         } else {
-            data
+            text = data
         }
-    } else {
-        str!()
     }
+
+    (media_type, text)
 }
 
 pub fn decode_url(input: String) -> String {
+    let input: String = input.replace("+", "%2B");
+
     form_urlencoded::parse(input.as_bytes())
         .map(|(key, val)| {
             [
@@ -200,11 +219,12 @@ pub fn file_url_to_fs_path(url: &str) -> String {
         fs_file_path = fs_file_path.replace("/", "\\");
     }
 
-    fs_file_path
+    // File paths should not be %-encoded
+    decode_url(fs_file_path)
 }
 
 pub fn retrieve_asset(
-    cache: &mut HashMap<String, String>,
+    cache: &mut HashMap<String, Vec<u8>>,
     client: &Client,
     parent_url: &str,
     url: &str,
@@ -216,10 +236,14 @@ pub fn retrieve_asset(
         return Ok((str!(), str!()));
     }
 
-    let cache_key = clean_url(&url);
-
     if is_data_url(&url) {
-        Ok((url.to_string(), url.to_string()))
+        if as_data_url {
+            Ok((url.to_string(), url.to_string()))
+        } else {
+            let (_media_type, text) = data_url_to_text(url);
+
+            Ok((text, url.to_string()))
+        }
     } else if is_file_url(&url) {
         // Check if parent_url is also file:///
         // (if not, then we don't embed the asset)
@@ -251,13 +275,25 @@ pub fn retrieve_asset(
             Ok((str!(), url.to_string()))
         }
     } else {
+        let cache_key: String = clean_url(&url);
+
         if cache.contains_key(&cache_key) {
-            // URL is in cache
+            // URL is in cache, we retrieve it
+            let data = cache.get(&cache_key).unwrap();
+
             if !opt_silent {
                 eprintln!("{} (from cache)", &url);
             }
-            let data = cache.get(&cache_key).unwrap();
-            Ok((data.to_string(), url.to_string()))
+
+            if as_data_url {
+                let url_fragment = get_url_fragment(url);
+                Ok((
+                    data_to_data_url(media_type, data, url, &url_fragment),
+                    url.to_string(),
+                ))
+            } else {
+                Ok((String::from_utf8_lossy(data).to_string(), url.to_string()))
+            }
         } else {
             // URL not in cache, we request it
             let mut response = client.get(url).send()?;
@@ -271,7 +307,7 @@ pub fn retrieve_asset(
                 }
             }
 
-            let new_cache_key = clean_url(&res_url);
+            let new_cache_key: String = clean_url(&res_url);
 
             if as_data_url {
                 // Convert response into a byte array
@@ -290,13 +326,17 @@ pub fn retrieve_asset(
                 };
                 let url_fragment = get_url_fragment(url);
                 let data_url = data_to_data_url(&media_type, &data, url, &url_fragment);
+
                 // Add to cache
-                cache.insert(new_cache_key, data_url.clone());
+                cache.insert(new_cache_key, data);
+
                 Ok((data_url, res_url))
             } else {
                 let content = response.text().unwrap();
+
                 // Add to cache
-                cache.insert(new_cache_key, content.clone());
+                cache.insert(new_cache_key, content.as_bytes().to_vec());
+
                 Ok((content, res_url))
             }
         }