server: handle multipart/mixed with an html or text subpart

Cargo.lock

@@ -36,6 +36,19 @@ dependencies = [
  "memchr",
 ]
 
+[[package]]
+name = "ammonia"
+version = "3.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "64e6d1c7838db705c9b756557ee27c384ce695a1c51a6fe528784cb1c6840170"
+dependencies = [
+ "html5ever 0.26.0",
+ "maplit",
+ "once_cell",
+ "tendril",
+ "url 2.4.1",
+]
+
 [[package]]
 name = "android-tzdata"
 version = "0.1.1"
@@ -1261,7 +1274,21 @@ checksum = "e5c13fb08e5d4dfc151ee5e88bae63f7773d61852f3bdc73c9f4b9e1bde03148"
 dependencies = [
  "log 0.4.20",
  "mac",
- "markup5ever",
+ "markup5ever 0.10.1",
+ "proc-macro2 1.0.66",
+ "quote 1.0.33",
+ "syn 1.0.109",
+]
+
+[[package]]
+name = "html5ever"
+version = "0.26.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bea68cab48b8459f17cf1c944c67ddc572d272d9f2b274140f223ecb1da4a3b7"
+dependencies = [
+ "log 0.4.20",
+ "mac",
+ "markup5ever 0.11.0",
  "proc-macro2 1.0.66",
  "quote 1.0.33",
  "syn 1.0.109",
@@ -1523,7 +1550,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1ea8e9c6e031377cff82ee3001dc8026cdf431ed4e2e6b51f98ab8c73484a358"
 dependencies = [
  "cssparser 0.27.2",
- "html5ever",
+ "html5ever 0.25.2",
  "matches",
  "selectors",
 ]
@@ -1637,6 +1664,12 @@ dependencies = [
  "quoted_printable",
 ]
 
+[[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"
@@ -1645,7 +1678,21 @@ checksum = "a24f40fb03852d1cdd84330cddcaf98e9ec08a7b7768e952fad3b4cf048ec8fd"
 dependencies = [
  "log 0.4.20",
  "phf 0.8.0",
- "phf_codegen",
+ "phf_codegen 0.8.0",
+ "string_cache",
+ "string_cache_codegen",
+ "tendril",
+]
+
+[[package]]
+name = "markup5ever"
+version = "0.11.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7a2629bb1404f3d34c2e921f21fd34ba00b206124c81f65c50b43b6aaefeb016"
+dependencies = [
+ "log 0.4.20",
+ "phf 0.10.1",
+ "phf_codegen 0.10.0",
  "string_cache",
  "string_cache_codegen",
  "tendril",
@@ -2081,6 +2128,16 @@ dependencies = [
  "phf_shared 0.8.0",
 ]
 
+[[package]]
+name = "phf_codegen"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4fb1c3a8bc4dd4e5cfce29b44ffc14bedd2ee294559a294e2a4d4c9e9a6a13cd"
+dependencies = [
+ "phf_generator 0.10.0",
+ "phf_shared 0.10.0",
+]
+
 [[package]]
 name = "phf_generator"
 version = "0.8.0"
@@ -2767,7 +2824,7 @@ dependencies = [
  "log 0.4.20",
  "matches",
  "phf 0.8.0",
- "phf_codegen",
+ "phf_codegen 0.8.0",
  "precomputed-hash",
  "servo_arc",
  "smallvec",
@@ -2836,6 +2893,7 @@ dependencies = [
 name = "server"
 version = "0.1.0"
 dependencies = [
+ "ammonia",
  "async-graphql",
  "async-graphql-rocket",
  "glog",

server/Cargo.toml

@@ -23,6 +23,7 @@ rocket_cors = "0.6.0"
 rayon = "1.8.0"
 memmap = "0.7.0"
 mailparse = "0.14.0"
+ammonia = "3.3.0"
 
 [dependencies.rocket_contrib]
 version = "0.4.11"

server/src/graphql.rs

@@ -8,7 +8,7 @@ use async_graphql::{
     Context, EmptyMutation, EmptySubscription, Error, FieldResult, Object, Schema, SimpleObject,
     Union,
 };
-use log::{info, warn};
+use log::{error, info, warn};
 use mailparse::{parse_mail, MailHeaderMap, ParsedMail};
 use memmap::MmapOptions;
 use notmuch::Notmuch;
@@ -58,6 +58,8 @@ pub struct Message {
     pub timestamp: Option<i64>,
     // The body contents
     pub body: Body,
+    // On disk location of message
+    pub path: String,
 }
 
 #[derive(Debug)]
@@ -241,18 +243,11 @@ impl QueryRoot {
                 .headers
                 .get_first_value("date")
                 .and_then(|d| mailparse::dateparse(&d).ok());
-            let body = m.get_body()?;
-            let body = match m.ctype.mimetype.as_str() {
-                "text/plain" => Body::PlainText(PlainText { text: body }),
-                "text/html" => Body::Html(Html { html: body }),
-                _ => {
-                    let msg = format!(
-                        "Unhandled body content type:\n{}",
-                        render_content_type_tree(&m)
-                    );
-                    warn!("{}", msg);
-                    Body::UnhandledContentType(UnhandledContentType { text: msg })
-                }
+            let body = match extract_body(&m)? {
+                Body::Html(Html { html }) => Body::Html(Html {
+                    html: ammonia::clean(&html),
+                }),
+                b => b,
             };
             messages.push(Message {
                 from,
@@ -261,6 +256,7 @@ impl QueryRoot {
                 subject,
                 timestamp,
                 body,
+                path,
             });
         }
         messages.reverse();
@@ -276,12 +272,92 @@ impl QueryRoot {
     }
 }
 
+fn extract_body(m: &ParsedMail) -> Result<Body, Error> {
+    let body = m.get_body()?;
+    let ret = match m.ctype.mimetype.as_str() {
+        "text/plain" => return Ok(Body::PlainText(PlainText { text: body })),
+        "text/html" => return Ok(Body::Html(Html { html: body })),
+        "multipart/mixed" => extract_mixed(m),
+        "multipart/alternative" => extract_alternative(m),
+        _ => extract_unhandled(m),
+    };
+    if let Err(err) = ret {
+        error!("Failed to extract body: {err:?}");
+        return Ok(extract_unhandled(m)?);
+    }
+    ret
+}
+
+fn extract_unhandled(m: &ParsedMail) -> Result<Body, Error> {
+    let msg = format!(
+        "Unhandled body content type:\n{}",
+        render_content_type_tree(m)
+    );
+    warn!("{}", msg);
+    Ok(Body::UnhandledContentType(UnhandledContentType {
+        text: msg,
+    }))
+}
+fn extract_alternative(m: &ParsedMail) -> Result<Body, Error> {
+    for sp in &m.subparts {
+        if sp.ctype.mimetype == "text/html" {
+            let body = sp.get_body()?;
+            return Ok(Body::Html(Html { html: body }));
+        }
+    }
+    for sp in &m.subparts {
+        if sp.ctype.mimetype == "text/plain" {
+            let body = sp.get_body()?;
+            return Ok(Body::PlainText(PlainText { text: body }));
+        }
+    }
+    Err("extract_alternative".into())
+}
+
+fn extract_mixed(m: &ParsedMail) -> Result<Body, Error> {
+    for sp in &m.subparts {
+        if sp.ctype.mimetype == "multipart/alternative" {
+            return extract_alternative(sp);
+        }
+    }
+    for sp in &m.subparts {
+        if sp.ctype.mimetype == "multipart/related" {
+            return extract_related(sp);
+        }
+    }
+    for sp in &m.subparts {
+        let body = sp.get_body()?;
+        match sp.ctype.mimetype.as_str() {
+            "text/plain" => return Ok(Body::PlainText(PlainText { text: body })),
+            "text/html" => return Ok(Body::Html(Html { html: body })),
+            _ => (),
+        }
+    }
+    Err("extract_mixed".into())
+}
+
+fn extract_related(m: &ParsedMail) -> Result<Body, Error> {
+    // TODO(wathiede): collect related things and change return type to new Body arm.
+    for sp in &m.subparts {
+        if sp.ctype.mimetype == "text/html" {
+            let body = sp.get_body()?;
+            return Ok(Body::Html(Html { html: body }));
+        }
+    }
+    for sp in &m.subparts {
+        if sp.ctype.mimetype == "text/plain" {
+            let body = sp.get_body()?;
+            return Ok(Body::PlainText(PlainText { text: body }));
+        }
+    }
+    Err("extract_related".into())
+}
+
 fn render_content_type_tree(m: &ParsedMail) -> String {
     const WIDTH: usize = 4;
     fn render_rec(m: &ParsedMail, depth: usize) -> String {
         let mut parts = Vec::new();
         let msg = format!("{} {}", "-".repeat(depth * WIDTH), m.ctype.mimetype);
-        println!("{msg}",);
         parts.push(msg);
         if !m.ctype.charset.is_empty() {
             parts.push(format!(

web/graphql/schema.json

@@ -290,6 +290,22 @@
                   "ofType": null
                 }
               }
+            },
+            {
+              "args": [],
+              "deprecationReason": null,
+              "description": null,
+              "isDeprecated": false,
+              "name": "path",
+              "type": {
+                "kind": "NON_NULL",
+                "name": null,
+                "ofType": {
+                  "kind": "SCALAR",
+                  "name": "String",
+                  "ofType": null
+                }
+              }
             }
           ],
           "inputFields": null,

web/graphql/show_thread.graphql

@@ -28,6 +28,7 @@ query ShowThreadQuery($threadId: String!) {
           contents
         }
       }
+      path
     }
   }
   tags {

web/index.html

@@ -124,12 +124,15 @@ input::placeholder, .input::placeholder{
   float: right;
 }
 /* Hide quoted emails */
+/*
 div[name="quote"],
 blockquote[type="cite"],
 .gmail_quote {
   background-color: red;
   display: none;
 }
+*/
+
 .desktop-main-content {
   display: grid;
   grid-template-columns: 12rem 1fr;