logiguard fork v3: full patch set on verified 8c74db0 tree

Includes prior-session patches (carry forward so the app compiles): - crates/gpui/build.rs: cross-compile manifest fix - crates/gpui/src/platform.rs: PlatformWindow::activate_with_token trait method - crates/gpui/src/window.rs: Window::activate_with_token public API - crates/gpui_linux/src/linux/wayland/window.rs: WaylandWindow::activate_with_token + activate() keyboard-serial fix Plus the focus-serial fix: - serial.rs: SerialKind::KeyboardEnter - client.rs: store wl_keyboard.enter serial; latest_serial_of() Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-14 01:52:12 +03:30
commit b9819977a5
3984 changed files with 1487015 additions and 0 deletions
--- a/crates/html_to_markdown/Cargo.toml
+++ b/crates/html_to_markdown/Cargo.toml
@@ -0,0 +1,26 @@
+[package]
+name = "html_to_markdown"
+version = "0.1.0"
+description = "Convert HTML to Markdown"
+repository = "https://github.com/zed-industries/zed"
+documentation = "https://docs.rs/html_to_markdown"
+keywords = ["html", "markdown", "html-to-markdown"]
+edition.workspace = true
+publish = true
+license = "Apache-2.0"
+
+[lints]
+workspace = true
+
+[lib]
+path = "src/html_to_markdown.rs"
+
+[dependencies]
+anyhow.workspace = true
+html5ever.workspace = true
+markup5ever_rcdom.workspace = true
+regex.workspace = true
+
+[dev-dependencies]
+indoc.workspace = true
+pretty_assertions.workspace = true
--- a/crates/html_to_markdown/LICENSE-APACHE
+++ b/crates/html_to_markdown/LICENSE-APACHE
@@ -0,0 +1 @@
+../../LICENSE-APACHE
--- a/crates/html_to_markdown/src/html_element.rs
+++ b/crates/html_to_markdown/src/html_element.rs
@@ -0,0 +1,83 @@
+use std::cell::RefCell;
+use std::collections::HashSet;
+use std::sync::OnceLock;
+
+use html5ever::Attribute;
+
+/// Returns a [`HashSet`] containing the HTML elements that are inline by default.
+///
+/// [MDN: List of "inline" elements](https://yari-demos.prod.mdn.mozit.cloud/en-US/docs/Web/HTML/Inline_elements)
+fn inline_elements() -> &'static HashSet<&'static str> {
+    static INLINE_ELEMENTS: OnceLock<HashSet<&str>> = OnceLock::new();
+    INLINE_ELEMENTS.get_or_init(|| {
+        HashSet::from_iter([
+            "a", "abbr", "acronym", "audio", "b", "bdi", "bdo", "big", "br", "button", "canvas",
+            "cite", "code", "data", "datalist", "del", "dfn", "em", "embed", "i", "iframe", "img",
+            "input", "ins", "kbd", "label", "map", "mark", "meter", "noscript", "object", "output",
+            "picture", "progress", "q", "ruby", "s", "samp", "script", "select", "slot", "small",
+            "span", "strong", "sub", "sup", "svg", "template", "textarea", "time", "tt", "u",
+            "var", "video", "wbr",
+        ])
+    })
+}
+
+#[derive(Debug, Clone)]
+pub struct HtmlElement {
+    tag: String,
+    pub(crate) attrs: RefCell<Vec<Attribute>>,
+}
+
+impl HtmlElement {
+    pub fn new(tag: String, attrs: RefCell<Vec<Attribute>>) -> Self {
+        Self { tag, attrs }
+    }
+
+    pub fn tag(&self) -> &str {
+        &self.tag
+    }
+
+    /// Returns whether this [`HtmlElement`] is an inline element.
+    pub fn is_inline(&self) -> bool {
+        inline_elements().contains(self.tag.as_str())
+    }
+
+    /// Returns the attribute with the specified name.
+    pub fn attr(&self, name: &str) -> Option<String> {
+        self.attrs
+            .borrow()
+            .iter()
+            .find(|attr| attr.name.local.to_string() == name)
+            .map(|attr| attr.value.to_string())
+    }
+
+    /// Returns the list of classes on this [`HtmlElement`].
+    pub fn classes(&self) -> Vec<String> {
+        self.attrs
+            .borrow()
+            .iter()
+            .find(|attr| attr.name.local.to_string() == "class")
+            .map(|attr| {
+                attr.value
+                    .split(' ')
+                    .map(|class| class.trim().to_string())
+                    .collect::<Vec<_>>()
+            })
+            .unwrap_or_default()
+    }
+
+    /// Returns whether this [`HtmlElement`] has the specified class.
+    pub fn has_class(&self, class: &str) -> bool {
+        self.has_any_classes(&[class])
+    }
+
+    /// Returns whether this [`HtmlElement`] has any of the specified classes.
+    pub fn has_any_classes(&self, classes: &[&str]) -> bool {
+        self.attrs.borrow().iter().any(|attr| {
+            attr.name.local.to_string() == "class"
+                && attr
+                    .value
+                    .split(' ')
+                    .any(|class| classes.contains(&class.trim()))
+        })
+    }
+}
--- a/crates/html_to_markdown/src/html_to_markdown.rs
+++ b/crates/html_to_markdown/src/html_to_markdown.rs
@@ -0,0 +1,46 @@
+//! Convert HTML to Markdown.
+
+mod html_element;
+pub mod markdown;
+mod markdown_writer;
+pub mod structure;
+
+use std::io::Read;
+
+use anyhow::{Context as _, Result};
+use html5ever::driver::ParseOpts;
+use html5ever::parse_document;
+use html5ever::tendril::TendrilSink;
+use html5ever::tree_builder::TreeBuilderOpts;
+use markup5ever_rcdom::RcDom;
+
+pub use crate::html_element::*;
+pub use crate::markdown_writer::*;
+
+/// Converts the provided HTML to Markdown.
+pub fn convert_html_to_markdown(html: impl Read, handlers: &mut [TagHandler]) -> Result<String> {
+    let dom = parse_html(html).context("failed to parse HTML")?;
+
+    let markdown_writer = MarkdownWriter::new();
+    let markdown = markdown_writer
+        .run(&dom.document, handlers)
+        .context("failed to convert HTML to Markdown")?;
+
+    Ok(markdown)
+}
+
+fn parse_html(mut html: impl Read) -> Result<RcDom> {
+    let parse_options = ParseOpts {
+        tree_builder: TreeBuilderOpts {
+            drop_doctype: true,
+            ..Default::default()
+        },
+        ..Default::default()
+    };
+    let dom = parse_document(RcDom::default(), parse_options)
+        .from_utf8()
+        .read_from(&mut html)
+        .context("failed to parse HTML document")?;
+
+    Ok(dom)
+}
--- a/crates/html_to_markdown/src/markdown.rs
+++ b/crates/html_to_markdown/src/markdown.rs
@@ -0,0 +1,276 @@
+use crate::html_element::HtmlElement;
+use crate::markdown_writer::{HandleTag, HandlerOutcome, MarkdownWriter, StartTagOutcome};
+
+pub struct WebpageChromeRemover;
+
+impl HandleTag for WebpageChromeRemover {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "head" | "script" | "style" | "nav")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        _writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "head" | "script" | "style" | "nav" => return StartTagOutcome::Skip,
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+}
+
+pub struct ParagraphHandler;
+
+impl HandleTag for ParagraphHandler {
+    fn should_handle(&self, _tag: &str) -> bool {
+        true
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        if tag.is_inline()
+            && writer.is_inside("p")
+            && let Some(parent) = writer.current_element_stack().iter().last()
+            && !(parent.is_inline()
+                || writer.markdown.ends_with(' ')
+                || writer.markdown.ends_with('\n'))
+        {
+            writer.push_str(" ");
+        }
+
+        if tag.tag() == "p" {
+            writer.push_blank_line()
+        }
+        StartTagOutcome::Continue
+    }
+}
+
+pub struct HeadingHandler;
+
+impl HandleTag for HeadingHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "h1" | "h2" | "h3" | "h4" | "h5" | "h6")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "h1" => writer.push_str("\n\n# "),
+            "h2" => writer.push_str("\n\n## "),
+            "h3" => writer.push_str("\n\n### "),
+            "h4" => writer.push_str("\n\n#### "),
+            "h5" => writer.push_str("\n\n##### "),
+            "h6" => writer.push_str("\n\n###### "),
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn handle_tag_end(&mut self, tag: &HtmlElement, writer: &mut MarkdownWriter) {
+        match tag.tag() {
+            "h1" | "h2" | "h3" | "h4" | "h5" | "h6" => writer.push_blank_line(),
+            _ => {}
+        }
+    }
+}
+
+pub struct ListHandler;
+
+impl HandleTag for ListHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "ul" | "ol" | "li")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "ul" | "ol" => writer.push_newline(),
+            "li" => writer.push_str("- "),
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn handle_tag_end(&mut self, tag: &HtmlElement, writer: &mut MarkdownWriter) {
+        match tag.tag() {
+            "ul" | "ol" => writer.push_newline(),
+            "li" => writer.push_newline(),
+            _ => {}
+        }
+    }
+}
+
+pub struct TableHandler {
+    /// The number of columns in the current `<table>`.
+    current_table_columns: usize,
+    is_first_th: bool,
+    is_first_td: bool,
+}
+
+impl TableHandler {
+    pub fn new() -> Self {
+        Self {
+            current_table_columns: 0,
+            is_first_th: true,
+            is_first_td: true,
+        }
+    }
+}
+
+impl Default for TableHandler {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl HandleTag for TableHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "table" | "thead" | "tbody" | "tr" | "th" | "td")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "thead" => writer.push_blank_line(),
+            "tr" => writer.push_newline(),
+            "th" => {
+                self.current_table_columns += 1;
+                if self.is_first_th {
+                    self.is_first_th = false;
+                } else {
+                    writer.push_str(" ");
+                }
+                writer.push_str("| ");
+            }
+            "td" => {
+                if self.is_first_td {
+                    self.is_first_td = false;
+                } else {
+                    writer.push_str(" ");
+                }
+                writer.push_str("| ");
+            }
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn handle_tag_end(&mut self, tag: &HtmlElement, writer: &mut MarkdownWriter) {
+        match tag.tag() {
+            "thead" => {
+                writer.push_newline();
+                for ix in 0..self.current_table_columns {
+                    if ix > 0 {
+                        writer.push_str(" ");
+                    }
+                    writer.push_str("| ---");
+                }
+                writer.push_str(" |");
+                self.is_first_th = true;
+            }
+            "tr" => {
+                writer.push_str(" |");
+                self.is_first_td = true;
+            }
+            "table" => {
+                self.current_table_columns = 0;
+            }
+            _ => {}
+        }
+    }
+}
+
+pub struct StyledTextHandler;
+
+impl HandleTag for StyledTextHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "strong" | "em")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "strong" => writer.push_str("**"),
+            "em" => writer.push_str("_"),
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn handle_tag_end(&mut self, tag: &HtmlElement, writer: &mut MarkdownWriter) {
+        match tag.tag() {
+            "strong" => writer.push_str("**"),
+            "em" => writer.push_str("_"),
+            _ => {}
+        }
+    }
+}
+
+pub struct CodeHandler;
+
+impl HandleTag for CodeHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "pre" | "code")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "code" => {
+                if !writer.is_inside("pre") {
+                    writer.push_str("`");
+                }
+            }
+            "pre" => writer.push_str("\n\n```\n"),
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn handle_tag_end(&mut self, tag: &HtmlElement, writer: &mut MarkdownWriter) {
+        match tag.tag() {
+            "code" => {
+                if !writer.is_inside("pre") {
+                    writer.push_str("`");
+                }
+            }
+            "pre" => writer.push_str("\n```\n"),
+            _ => {}
+        }
+    }
+
+    fn handle_text(&mut self, text: &str, writer: &mut MarkdownWriter) -> HandlerOutcome {
+        if writer.is_inside("pre") {
+            writer.push_str(text);
+            return HandlerOutcome::Handled;
+        }
+
+        HandlerOutcome::NoOp
+    }
+}
--- a/crates/html_to_markdown/src/markdown_writer.rs
+++ b/crates/html_to_markdown/src/markdown_writer.rs
@@ -0,0 +1,199 @@
+use std::collections::VecDeque;
+use std::rc::Rc;
+use std::{cell::RefCell, sync::LazyLock};
+
+use anyhow::Result;
+use markup5ever_rcdom::{Handle, NodeData};
+use regex::Regex;
+
+use crate::html_element::HtmlElement;
+
+fn empty_line_regex() -> &'static Regex {
+    static REGEX: LazyLock<Regex> =
+        LazyLock::new(|| Regex::new(r"^\s*$").expect("Failed to create empty_line_regex"));
+    &REGEX
+}
+
+fn more_than_three_newlines_regex() -> &'static Regex {
+    static REGEX: LazyLock<Regex> = LazyLock::new(|| Regex::new(r"\n{3,}").unwrap());
+    &REGEX
+}
+
+pub enum StartTagOutcome {
+    Continue,
+    Skip,
+}
+
+pub type TagHandler = Rc<RefCell<dyn HandleTag>>;
+
+pub struct MarkdownWriter {
+    current_element_stack: VecDeque<HtmlElement>,
+    pub(crate) markdown: String,
+}
+
+impl Default for MarkdownWriter {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl MarkdownWriter {
+    pub fn new() -> Self {
+        Self {
+            current_element_stack: VecDeque::new(),
+            markdown: String::new(),
+        }
+    }
+
+    pub fn current_element_stack(&self) -> &VecDeque<HtmlElement> {
+        &self.current_element_stack
+    }
+
+    pub fn is_inside(&self, tag: &str) -> bool {
+        self.current_element_stack
+            .iter()
+            .any(|parent_element| parent_element.tag() == tag)
+    }
+
+    /// Appends the given string slice onto the end of the Markdown output.
+    pub fn push_str(&mut self, str: &str) {
+        self.markdown.push_str(str);
+    }
+
+    /// Appends a newline to the end of the Markdown output.
+    pub fn push_newline(&mut self) {
+        self.push_str("\n");
+    }
+
+    /// Appends a blank line to the end of the Markdown output.
+    pub fn push_blank_line(&mut self) {
+        self.push_str("\n\n");
+    }
+
+    pub fn run(mut self, root_node: &Handle, handlers: &mut [TagHandler]) -> Result<String> {
+        self.visit_node(root_node, handlers)?;
+        Ok(Self::prettify_markdown(self.markdown))
+    }
+
+    fn prettify_markdown(markdown: String) -> String {
+        let markdown = empty_line_regex().replace_all(&markdown, "");
+        let markdown = more_than_three_newlines_regex().replace_all(&markdown, "\n\n");
+
+        markdown.trim().to_string()
+    }
+
+    fn visit_node(&mut self, node: &Handle, handlers: &mut [TagHandler]) -> Result<()> {
+        let mut current_element = None;
+
+        match node.data {
+            NodeData::Document
+            | NodeData::Doctype { .. }
+            | NodeData::ProcessingInstruction { .. }
+            | NodeData::Comment { .. } => {
+                // Currently left unimplemented, as we're not interested in this data
+                // at this time.
+            }
+            NodeData::Element {
+                ref name,
+                ref attrs,
+                ..
+            } => {
+                let tag_name = name.local.to_string();
+                if !tag_name.is_empty() {
+                    current_element = Some(HtmlElement::new(tag_name, attrs.clone()));
+                }
+            }
+            NodeData::Text { ref contents } => {
+                let text = contents.borrow().to_string();
+                self.visit_text(text, handlers)?;
+            }
+        }
+
+        if let Some(current_element) = current_element.as_ref() {
+            match self.start_tag(current_element, handlers) {
+                StartTagOutcome::Continue => {}
+                StartTagOutcome::Skip => return Ok(()),
+            }
+
+            self.current_element_stack
+                .push_back(current_element.clone());
+        }
+
+        if self.current_element_stack.len() < 200 {
+            for child in node.children.borrow().iter() {
+                self.visit_node(child, handlers)?;
+            }
+        }
+
+        if let Some(current_element) = current_element {
+            self.current_element_stack.pop_back();
+            self.end_tag(&current_element, handlers);
+        }
+
+        Ok(())
+    }
+
+    fn start_tag(&mut self, tag: &HtmlElement, handlers: &mut [TagHandler]) -> StartTagOutcome {
+        for handler in handlers {
+            if handler.borrow().should_handle(tag.tag()) {
+                match handler.borrow_mut().handle_tag_start(tag, self) {
+                    StartTagOutcome::Continue => {}
+                    StartTagOutcome::Skip => return StartTagOutcome::Skip,
+                }
+            }
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn end_tag(&mut self, tag: &HtmlElement, handlers: &mut [TagHandler]) {
+        for handler in handlers {
+            if handler.borrow().should_handle(tag.tag()) {
+                handler.borrow_mut().handle_tag_end(tag, self);
+            }
+        }
+    }
+
+    fn visit_text(&mut self, text: String, handlers: &mut [TagHandler]) -> Result<()> {
+        for handler in handlers {
+            match handler.borrow_mut().handle_text(&text, self) {
+                HandlerOutcome::Handled => return Ok(()),
+                HandlerOutcome::NoOp => {}
+            }
+        }
+
+        let text = text
+            .trim_matches(|char| char == '\n' || char == '\r' || char == '\t')
+            .replace('\n', " ");
+
+        self.push_str(&text);
+
+        Ok(())
+    }
+}
+
+pub enum HandlerOutcome {
+    Handled,
+    NoOp,
+}
+
+pub trait HandleTag {
+    /// Returns whether this handler should handle the given tag.
+    fn should_handle(&self, tag: &str) -> bool;
+
+    /// Handles the start of the given tag.
+    fn handle_tag_start(
+        &mut self,
+        _tag: &HtmlElement,
+        _writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        StartTagOutcome::Continue
+    }
+
+    /// Handles the end of the given tag.
+    fn handle_tag_end(&mut self, _tag: &HtmlElement, _writer: &mut MarkdownWriter) {}
+
+    fn handle_text(&mut self, _text: &str, _writer: &mut MarkdownWriter) -> HandlerOutcome {
+        HandlerOutcome::NoOp
+    }
+}
--- a/crates/html_to_markdown/src/structure.rs
+++ b/crates/html_to_markdown/src/structure.rs
@@ -0,0 +1 @@
+pub mod wikipedia;
--- a/crates/html_to_markdown/src/structure/wikipedia.rs
+++ b/crates/html_to_markdown/src/structure/wikipedia.rs
@@ -0,0 +1,181 @@
+use crate::HandleTag;
+use crate::html_element::HtmlElement;
+use crate::markdown_writer::{HandlerOutcome, MarkdownWriter, StartTagOutcome};
+
+pub struct WikipediaChromeRemover;
+
+impl HandleTag for WikipediaChromeRemover {
+    fn should_handle(&self, _tag: &str) -> bool {
+        true
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        _writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "head" | "script" | "style" | "nav" => return StartTagOutcome::Skip,
+            "sup" => {
+                if tag.has_class("reference") {
+                    return StartTagOutcome::Skip;
+                }
+            }
+            "div" | "span" | "a" => {
+                if tag.attr("id").as_deref() == Some("p-lang-btn") {
+                    return StartTagOutcome::Skip;
+                }
+
+                if tag.attr("id").as_deref() == Some("p-search") {
+                    return StartTagOutcome::Skip;
+                }
+
+                let classes_to_skip = ["noprint", "mw-editsection", "mw-jump-link"];
+                if tag.has_any_classes(&classes_to_skip) {
+                    return StartTagOutcome::Skip;
+                }
+            }
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+}
+
+pub struct WikipediaInfoboxHandler;
+
+impl HandleTag for WikipediaInfoboxHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        tag == "table"
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        _writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        if tag.tag() == "table" && tag.has_class("infobox") {
+            return StartTagOutcome::Skip;
+        }
+
+        StartTagOutcome::Continue
+    }
+}
+
+pub struct WikipediaCodeHandler {
+    language: Option<String>,
+}
+
+impl WikipediaCodeHandler {
+    pub fn new() -> Self {
+        Self { language: None }
+    }
+}
+
+impl Default for WikipediaCodeHandler {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl HandleTag for WikipediaCodeHandler {
+    fn should_handle(&self, tag: &str) -> bool {
+        matches!(tag, "div" | "pre" | "code")
+    }
+
+    fn handle_tag_start(
+        &mut self,
+        tag: &HtmlElement,
+        writer: &mut MarkdownWriter,
+    ) -> StartTagOutcome {
+        match tag.tag() {
+            "code" => {
+                if !writer.is_inside("pre") {
+                    writer.push_str("`");
+                }
+            }
+            "div" => {
+                let classes = tag.classes();
+                self.language = classes.iter().find_map(|class| {
+                    if let Some((_, language)) = class.split_once("mw-highlight-lang-") {
+                        Some(language.trim().to_owned())
+                    } else {
+                        None
+                    }
+                });
+            }
+            "pre" => {
+                writer.push_blank_line();
+                writer.push_str("```");
+                if let Some(language) = self.language.take() {
+                    writer.push_str(&language);
+                }
+                writer.push_newline();
+            }
+            _ => {}
+        }
+
+        StartTagOutcome::Continue
+    }
+
+    fn handle_tag_end(&mut self, tag: &HtmlElement, writer: &mut MarkdownWriter) {
+        match tag.tag() {
+            "code" => {
+                if !writer.is_inside("pre") {
+                    writer.push_str("`");
+                }
+            }
+            "pre" => writer.push_str("\n```\n"),
+            _ => {}
+        }
+    }
+
+    fn handle_text(&mut self, text: &str, writer: &mut MarkdownWriter) -> HandlerOutcome {
+        if writer.is_inside("pre") {
+            writer.push_str(text);
+            return HandlerOutcome::Handled;
+        }
+
+        HandlerOutcome::NoOp
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use std::cell::RefCell;
+    use std::rc::Rc;
+
+    use indoc::indoc;
+    use pretty_assertions::assert_eq;
+
+    use crate::{TagHandler, convert_html_to_markdown, markdown};
+
+    use super::*;
+
+    fn wikipedia_handlers() -> Vec<TagHandler> {
+        vec![
+            Rc::new(RefCell::new(markdown::ParagraphHandler)),
+            Rc::new(RefCell::new(markdown::HeadingHandler)),
+            Rc::new(RefCell::new(markdown::ListHandler)),
+            Rc::new(RefCell::new(markdown::StyledTextHandler)),
+            Rc::new(RefCell::new(WikipediaChromeRemover)),
+        ]
+    }
+
+    #[test]
+    fn test_citation_references_get_removed() {
+        let html = indoc! {r##"
+            <p>Rust began as a personal project in 2006 by <a href="/wiki/Mozilla" title="Mozilla">Mozilla</a> Research employee Graydon Hoare.<sup id="cite_ref-MITTechReview_23-0" class="reference"><a href="#cite_note-MITTechReview-23">[20]</a></sup> Mozilla began sponsoring the project in 2009 as a part of the ongoing development of an experimental <a href="/wiki/Browser_engine" title="Browser engine">browser engine</a> called <a href="/wiki/Servo_(software)" title="Servo (software)">Servo</a>,<sup id="cite_ref-infoq2012_24-0" class="reference"><a href="#cite_note-infoq2012-24">[21]</a></sup> which was officially announced by Mozilla in 2010.<sup id="cite_ref-MattAsay_25-0" class="reference"><a href="#cite_note-MattAsay-25">[22]</a></sup><sup id="cite_ref-26" class="reference"><a href="#cite_note-26">[23]</a></sup> Rust's memory and ownership system was influenced by <a href="/wiki/Region-based_memory_management" title="Region-based memory management">region-based memory management</a> in languages such as <a href="/wiki/Cyclone_(programming_language)" title="Cyclone (programming language)">Cyclone</a> and ML Kit.<sup id="cite_ref-influences_8-13" class="reference"><a href="#cite_note-influences-8">[5]</a></sup>
+            </p>
+        "##};
+        let expected = indoc! {"
+            Rust began as a personal project in 2006 by Mozilla Research employee Graydon Hoare.  Mozilla began sponsoring the project in 2009 as a part of the ongoing development of an experimental browser engine called Servo,  which was officially announced by Mozilla in 2010.  Rust's memory and ownership system was influenced by region-based memory management in languages such as Cyclone and ML Kit.
+        "}
+        .trim();
+
+        assert_eq!(
+            convert_html_to_markdown(html.as_bytes(), &mut wikipedia_handlers()).unwrap(),
+            expected
+        )
+    }
+}