Refactor & prepare for matrix HTML parsing

Cargo.lock

@@ -584,6 +584,16 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "futf"
+version = "0.1.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7c9c1ce3fa9336301af935ab852c437817d14cd33690446569392e65170aac3b"
+dependencies = [
+ "mac",
+ "new_debug_unreachable",
+]
+
 [[package]]
 name = "futures"
 version = "0.3.16"
@@ -857,6 +867,20 @@ dependencies = [
  "utf8-width",
 ]
 
+[[package]]
+name = "html5ever"
+version = "0.25.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "aafcf38a1a36118242d29b92e1b08ef84e67e4a5ed06e0a80be20e6a32bfed6b"
+dependencies = [
+ "log",
+ "mac",
+ "markup5ever",
+ "proc-macro2 1.0.28",
+ "quote 1.0.9",
+ "syn 1.0.74",
+]
+
 [[package]]
 name = "http"
 version = "0.2.4"
@@ -1108,12 +1132,32 @@ dependencies = [
  "hashbrown",
 ]
 
+[[package]]
+name = "mac"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c41e0c4fef86961ac6d6f8a82609f55f31b05e4fce149ac5710e439df7619ba4"
+
 [[package]]
 name = "maplit"
 version = "1.0.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3e2e65a1a2e43cfcb47a895c4c8b10d1f4a61097f9f254f183aee60cad9c651d"
 
+[[package]]
+name = "markup5ever"
+version = "0.10.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a24f40fb03852d1cdd84330cddcaf98e9ec08a7b7768e952fad3b4cf048ec8fd"
+dependencies = [
+ "log",
+ "phf",
+ "phf_codegen",
+ "string_cache",
+ "string_cache_codegen",
+ "tendril",
+]
+
 [[package]]
 name = "matches"
 version = "0.1.8"
@@ -1318,6 +1362,12 @@ dependencies = [
  "tempfile",
 ]
 
+[[package]]
+name = "new_debug_unreachable"
+version = "1.0.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e4a24736216ec316047a1fc4252e27dabb04218aa4a3f37c6e7ddbf1f9782b54"
+
 [[package]]
 name = "ntapi"
 version = "0.3.6"
@@ -1489,6 +1539,44 @@ version = "2.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d4fd5641d01c8f18a23da7b6fe29298ff4b55afcccdf78973b24cf3175fee32e"
 
+[[package]]
+name = "phf"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3dfb61232e34fcb633f43d12c58f83c1df82962dcdfa565a4e866ffc17dafe12"
+dependencies = [
+ "phf_shared",
+]
+
+[[package]]
+name = "phf_codegen"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cbffee61585b0411840d3ece935cce9cb6321f01c45477d30066498cd5e1a815"
+dependencies = [
+ "phf_generator",
+ "phf_shared",
+]
+
+[[package]]
+name = "phf_generator"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "17367f0cc86f2d25802b2c26ee58a7b23faeccf78a396094c13dced0d0182526"
+dependencies = [
+ "phf_shared",
+ "rand 0.7.3",
+]
+
+[[package]]
+name = "phf_shared"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c00cf8b9eafe68dde5e9eaa2cef8ee84a9336a47d566ec55ca16589633b65af7"
+dependencies = [
+ "siphasher",
+]
+
 [[package]]
 name = "phoebe"
 version = "0.1.0"
@@ -1497,6 +1585,7 @@ dependencies = [
  "discord_message_format",
  "env_logger",
  "html-escape",
+ "html5ever",
  "log",
  "matrix-sdk",
  "serde",
@@ -1596,6 +1685,12 @@ version = "0.2.10"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ac74c624d6b2d21f425f752262f42188365d7b8ff1aff74c82e45136510a4857"
 
+[[package]]
+name = "precomputed-hash"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "925383efa346730478fb4838dbe9137d2a47675ad789c546d150a6e1dd4ab31c"
+
 [[package]]
 name = "proc-macro-crate"
 version = "1.0.0"
@@ -1675,6 +1770,7 @@ dependencies = [
  "rand_chacha 0.2.2",
  "rand_core 0.5.1",
  "rand_hc 0.2.0",
+ "rand_pcg",
 ]
 
 [[package]]
@@ -1745,6 +1841,15 @@ dependencies = [
  "rand_core 0.6.3",
 ]
 
+[[package]]
+name = "rand_pcg"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "16abd0c1b639e9eb4d7c50c0b8100b0d0f849be2349829c740fe8e6eb4816429"
+dependencies = [
+ "rand_core 0.5.1",
+]
+
 [[package]]
 name = "rayon"
 version = "1.5.1"
@@ -2313,6 +2418,12 @@ version = "1.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c19772be3c4dd2ceaacf03cb41d5885f2a02c4d8804884918e3a258480803335"
 
+[[package]]
+name = "siphasher"
+version = "0.3.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "533494a8f9b724d33625ab53c6c4800f7cc445895924a8ef649222dcb76e938b"
+
 [[package]]
 name = "slab"
 version = "0.4.4"
@@ -2424,6 +2535,31 @@ version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "213701ba3370744dcd1a12960caa4843b3d68b4d1c0a5d575e0d65b2ee9d16c0"
 
+[[package]]
+name = "string_cache"
+version = "0.8.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8ddb1139b5353f96e429e1a5e19fbaf663bddedaa06d1dbd49f82e352601209a"
+dependencies = [
+ "lazy_static",
+ "new_debug_unreachable",
+ "phf_shared",
+ "precomputed-hash",
+ "serde",
+]
+
+[[package]]
+name = "string_cache_codegen"
+version = "0.5.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f24c8e5e19d22a726626f1a5e16fe15b132dcf21d10177fa5a45ce7962996b97"
+dependencies = [
+ "phf_generator",
+ "phf_shared",
+ "proc-macro2 1.0.28",
+ "quote 1.0.9",
+]
+
 [[package]]
 name = "subtle"
 version = "2.4.1"
@@ -2478,6 +2614,17 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "tendril"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a9ef557cb397a4f0a5a3a628f06515f78563f2209e64d47055d9dc6052bf5e33"
+dependencies = [
+ "futf",
+ "mac",
+ "utf-8",
+]
+
 [[package]]
 name = "termcolor"
 version = "1.1.2"

Cargo.toml

@@ -14,6 +14,7 @@ url = "2.2.2"
 log = "0.4.14"
 env_logger = "0.9.0"
 html-escape = "0.2.9"
+html5ever = "0.25.1"
 
 [dependencies.serenity]
 version = "0.10.9"

src/matrix.rs

@@ -20,7 +20,7 @@ use tokio::sync::mpsc;
 use url::Url;
 
 use crate::{
-    message_ast::{convert_plain, format_discord, format_matrix},
+    message_ast::{convert_matrix, convert_plain, format_discord, format_matrix},
     messages::{MessageAuthor, MessageReference, SentMessage},
 };
 
@@ -87,14 +87,14 @@ impl EventHandler for MatrixHandler {
 
             match message_type {
                 MessageType::Text(text) => {
-                    let content = if let Some(_html) = text
+                    let content = if let Some(html) = text
                         .formatted
                         .as_ref()
                         .filter(|f| f.format == MessageFormat::Html)
                         .map(|f| &f.body)
                     {
                         // TODO: Parse html_body into MessageContent AST
-                        convert_plain(&text.body)
+                        convert_matrix(html)
                     } else {
                         convert_plain(&text.body)
                     };
@@ -106,7 +106,7 @@ impl EventHandler for MatrixHandler {
                             author: MessageAuthor {
                                 display_name: sender
                                     .display_name()
-                                    .unwrap_or(sender.name())
+                                    .unwrap_or_else(|| sender.name())
                                     .to_string(),
                             },
                         });

src/message_ast/convert_discord.rs

@@ -1,7 +1,7 @@
-use super::{MessageContent, Styled};
+use super::{MessageComponent, MessageContent};
 use discord_message_format::DiscordComponent;
 
-impl<'a> From<&DiscordComponent<'a>> for Styled {
+impl<'a> From<&DiscordComponent<'a>> for MessageComponent {
     fn from(discord: &DiscordComponent<'a>) -> Self {
         match discord {
             DiscordComponent::Plain(text) => Self::Plain(text.to_string()),
@@ -38,38 +38,38 @@ impl<'a> From<&DiscordComponent<'a>> for Styled {
 pub fn convert_discord(discord_message: &[DiscordComponent<'_>]) -> MessageContent {
     discord_message
         .iter()
-        .map(Styled::from)
+        .map(MessageComponent::from)
         .collect::<MessageContent>()
 }
 
-pub fn format_discord(message_content: &MessageContent) -> String {
+pub fn format_discord(message_content: &[MessageComponent]) -> String {
     message_content
         .iter()
         .map(|component| match component {
-            Styled::Plain(text) => text.to_string(), // TODO: Escape
+            MessageComponent::Plain(text) => text.to_string(), // TODO: Escape
 
-            Styled::Link { target, .. } => target.to_string(), // TODO: Link text
+            MessageComponent::Link { target, .. } => target.to_string(), // TODO: Link text
 
-            Styled::Italic(inner) => format!("*{}*", format_discord(inner)),
+            MessageComponent::Italic(inner) => format!("*{}*", format_discord(inner)),
-            Styled::Bold(inner) => format!("**{}**", format_discord(inner)),
+            MessageComponent::Bold(inner) => format!("**{}**", format_discord(inner)),
-            Styled::Strikethrough(inner) => format!("~~{}~~", format_discord(inner)),
+            MessageComponent::Strikethrough(inner) => format!("~~{}~~", format_discord(inner)),
-            Styled::Underline(inner) => format!("__{}__", format_discord(inner)),
+            MessageComponent::Underline(inner) => format!("__{}__", format_discord(inner)),
 
-            Styled::Code(code) => format!("`{}`", code), // TODO: Double-grave delimiting when code contains '`'
+            MessageComponent::Code(code) => format!("`{}`", code), // TODO: Double-grave delimiting when code contains '`'
-            Styled::CodeBlock { lang, source } => {
+            MessageComponent::CodeBlock { lang, source } => {
                 format!(
                     "```{}\n{}\n```",
                     lang.as_ref()
                         .map(|s| s.to_string())
-                        .unwrap_or("".to_string()),
+                        .unwrap_or_else(|| "".to_string()),
                     source.to_string()
                 )
             }
 
-            Styled::Spoiler { content, .. } => format!("||{}||", format_discord(content)), // TODO: Spoiler reason
+            MessageComponent::Spoiler { content, .. } => format!("||{}||", format_discord(content)), // TODO: Spoiler reason
 
-            Styled::HardBreak => "\n".to_string(),
+            MessageComponent::HardBreak => "\n".to_string(),
-            Styled::BlockQuote(inner) => format!("> {}", format_discord(inner)),
+            MessageComponent::BlockQuote(inner) => format!("> {}", format_discord(inner)),
         })
         .collect()
 }

src/message_ast/convert_matrix.rs

@@ -1,27 +1,31 @@
-use super::{MessageContent, Styled};
+use super::{MessageComponent, MessageContent};
 
 pub fn convert_matrix(message: &str) -> MessageContent {
     todo!();
 }
 
-pub fn format_matrix(message_content: &MessageContent) -> String {
+pub fn format_matrix(message_content: &[MessageComponent]) -> String {
     message_content
         .iter()
         .map(|component| match component {
-            Styled::Plain(text) => html_escape::encode_text(text).to_string(),
+            MessageComponent::Plain(text) => html_escape::encode_text(text).to_string(),
-            Styled::Link { target, text } => format!(
+            MessageComponent::Link { target, text } => format!(
                 r#"<a href="{}">{}</a>"#,
                 html_escape::encode_quoted_attribute(target),
                 format_matrix(text)
             ),
 
-            Styled::Italic(inner) => format!("<em>{}</em>", format_matrix(inner)),
+            MessageComponent::Italic(inner) => format!("<em>{}</em>", format_matrix(inner)),
-            Styled::Bold(inner) => format!("<strong>{}</strong>", format_matrix(inner)),
+            MessageComponent::Bold(inner) => format!("<strong>{}</strong>", format_matrix(inner)),
-            Styled::Strikethrough(inner) => format!("<del>{}</del>", format_matrix(inner)),
+            MessageComponent::Strikethrough(inner) => {
-            Styled::Underline(inner) => format!("<u>{}</u>", format_matrix(inner)),
+                format!("<del>{}</del>", format_matrix(inner))
+            }
+            MessageComponent::Underline(inner) => format!("<u>{}</u>", format_matrix(inner)),
 
-            Styled::Code(code) => format!("<code>{}</code>", html_escape::encode_text(code)),
+            MessageComponent::Code(code) => {
-            Styled::CodeBlock { lang, source } => {
+                format!("<code>{}</code>", html_escape::encode_text(code))
+            }
+            MessageComponent::CodeBlock { lang, source } => {
                 format!(
                     r#"<pre><code{}>{}</code></pre>"#,
                     lang.as_ref()
@@ -29,21 +33,21 @@ pub fn format_matrix(message_content: &MessageContent) -> String {
                             r#" class="language-{}""#,
                             html_escape::encode_quoted_attribute(lang)
                         ))
-                        .unwrap_or("".to_string()),
+                        .unwrap_or_else(|| "".to_string()),
                     source,
                 )
             }
 
-            Styled::Spoiler { reason, content } => format!(
+            MessageComponent::Spoiler { reason, content } => format!(
                 "<span data-mx-spoiler{}>{}</span>",
                 reason
                     .as_ref()
                     .map(|reason| format!(r#"="{}""#, html_escape::encode_quoted_attribute(reason)))
-                    .unwrap_or("".to_string()),
+                    .unwrap_or_else(|| "".to_string()),
                 format_matrix(content)
             ),
-            Styled::HardBreak => "<br>".to_string(),
+            MessageComponent::HardBreak => "<br>".to_string(),
-            Styled::BlockQuote(inner) => {
+            MessageComponent::BlockQuote(inner) => {
                 format!("<blockquote>{}</blockquote>", format_matrix(inner))
             }
         })

src/message_ast/convert_plain.rs

@@ -1,5 +1,5 @@
-use super::{MessageContent, Styled};
+use super::{MessageComponent, MessageContent};
 
 pub fn convert_plain(message: &str) -> MessageContent {
-    vec![Styled::Plain(message.to_string())]
+    vec![MessageComponent::Plain(message.to_string())]
 }

src/message_ast/mod.rs

@@ -2,9 +2,9 @@ mod convert_discord;
 mod convert_matrix;
 mod convert_plain;
 
-pub type MessageContent = Vec<Styled>;
+pub type MessageContent = Vec<MessageComponent>;
 
-pub enum Styled {
+pub enum MessageComponent {
     Plain(String),
     Link {
         target: String,