0.5.0: Parse SVG size

Cargo.lock

@@ -2,6 +2,17 @@
 # It is not intended for manual editing.
 version = 3
 
+[[package]]
+name = "ahash"
+version = "0.7.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fcb51a0695d8f838b1ee009b3fbf66bda078cd64590202a864a8f3e8c4315c47"
+dependencies = [
+ "getrandom 0.2.8",
+ "once_cell",
+ "version_check",
+]
+
 [[package]]
 name = "aho-corasick"
 version = "0.7.19"
@@ -566,6 +577,9 @@ name = "hashbrown"
 version = "0.12.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8a9ee70c43aaf417c914396645a0fa852624801b24ebb7ae78fe8272889ac888"
+dependencies = [
+ "ahash",
+]
 
 [[package]]
 name = "heck"
@@ -792,6 +806,12 @@ version = "1.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
 
+[[package]]
+name = "lazycell"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
+
 [[package]]
 name = "libc"
 version = "0.2.137"
@@ -826,6 +846,25 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "lol_html"
+version = "0.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b16d7d17063591adf0017a068a8ce2ed29a4f5402b9e6cac01888ff9c0527ff7"
+dependencies = [
+ "bitflags",
+ "cfg-if",
+ "cssparser",
+ "encoding_rs",
+ "hashbrown",
+ "lazy_static",
+ "lazycell",
+ "memchr",
+ "safemem",
+ "selectors",
+ "thiserror",
+]
+
 [[package]]
 name = "mac"
 version = "0.1.1"
@@ -1453,6 +1492,12 @@ version = "1.0.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "4501abdff3ae82a1c1b477a17252eb69cee9e66eb915c1abaa4f44d873df9f09"
 
+[[package]]
+name = "safemem"
+version = "0.3.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ef703b7cb59335eae2eb93ceb664c0eb7ea6bf567079d843e09420219668e072"
+
 [[package]]
 name = "schannel"
 version = "0.1.20"
@@ -1638,7 +1683,7 @@ checksum = "7bd3e3206899af3f8b12af284fafc038cc1dc2b41d1b89dd17297221c5d225de"
 
 [[package]]
 name = "site_icons"
-version = "0.4.13"
+version = "0.5.0"
 dependencies = [
  "byteorder",
  "clap",
@@ -1648,6 +1693,7 @@ dependencies = [
  "html5ever",
  "itertools",
  "log",
+ "lol_html",
  "mime_4",
  "once_cell",
  "percent-encoding",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "site_icons"
-version = "0.4.13"
+version = "0.5.0"
 authors = ["Sam Denty <sam@samdenty.com>"]
 edition = "2018"
 license = "GPL-3.0"
@@ -35,6 +35,7 @@ serde = { version = "1.0", features = ["derive", "rc"] }
 serde_json = "1.0"
 futures = "0.3.25"
 tldextract = "0.6.0"
+lol_html = "0.3.2"
 
 [target.'cfg(target_arch = "wasm32")'.dependencies]
 reqwest = { package = "reqwest-wasm", version = "0.11.16", features = [

src/icon_info.rs

@@ -11,9 +11,8 @@ use std::{
   io,
 };
 
-#[derive(Serialize, Deserialize)]
-#[serde(rename_all = "lowercase")]
 enum IconKind {
+  SVG,
   PNG,
   JPEG,
   ICO,
@@ -28,7 +27,7 @@ pub enum IconInfo {
   JPEG { size: IconSize },
   ICO { sizes: IconSizes },
   GIF { size: IconSize },
-  SVG,
+  SVG { size: Option<IconSize> },
 }
 
 impl IconInfo {
@@ -40,6 +39,14 @@ impl IconInfo {
     reader.read_exact(&mut header).await?;
 
     match (kind, &header) {
+      (Some(IconKind::SVG), bytes) => {
+        let size = get_svg_size(bytes, reader).await?;
+        Ok(IconInfo::SVG { size })
+      }
+      (_, &[0x60, byte_two]) => {
+        let size = get_svg_size(&[0x60, byte_two], reader).await?;
+        Ok(IconInfo::SVG { size })
+      }
       (Some(IconKind::PNG), _) | (_, b"\x89P") => {
         let size = get_png_size(reader).await?;
         Ok(IconInfo::PNG { size })
@@ -135,7 +142,25 @@ impl IconInfo {
         Some(IconKind::ICO)
       }
 
-      (mime::IMAGE, mime::SVG) | (mime::TEXT, mime::PLAIN) => return Ok(IconInfo::SVG),
+      (mime::IMAGE, mime::GIF) => {
+        if let Some(sizes) = sizes {
+          return Ok(IconInfo::GIF {
+            size: *sizes.largest(),
+          });
+        }
+
+        Some(IconKind::GIF)
+      }
+
+      (mime::IMAGE, mime::SVG) | (mime::TEXT, mime::PLAIN) => {
+        if let Some(sizes) = sizes {
+          return Ok(IconInfo::SVG {
+            size: Some(*sizes.largest()),
+          });
+        }
+
+        Some(IconKind::SVG)
+      }
 
       _ => None,
     };
@@ -147,7 +172,7 @@ impl IconInfo {
     match self {
       IconInfo::ICO { sizes } => Some(sizes.largest()),
       IconInfo::PNG { size } | IconInfo::JPEG { size } | IconInfo::GIF { size } => Some(size),
-      IconInfo::SVG => None,
+      IconInfo::SVG { size } => size.as_ref(),
     }
   }
 
@@ -157,7 +182,7 @@ impl IconInfo {
       IconInfo::PNG { size } | IconInfo::JPEG { size } | IconInfo::GIF { size } => {
         Some((*size).into())
       }
-      IconInfo::SVG => None,
+      IconInfo::SVG { size } => size.map(|size| size.into()),
     }
   }
 
@@ -167,7 +192,7 @@ impl IconInfo {
       IconInfo::JPEG { .. } => "image/jpeg",
       IconInfo::ICO { .. } => "image/x-icon",
       IconInfo::GIF { .. } => "image/gif",
-      IconInfo::SVG => "image/svg+xml",
+      IconInfo::SVG { .. } => "image/svg+xml",
     }
   }
 }
@@ -179,7 +204,17 @@ impl Display for IconInfo {
       IconInfo::JPEG { size } => write!(f, "jpeg {}", size),
       IconInfo::GIF { size } => write!(f, "gif {}", size),
       IconInfo::ICO { sizes } => write!(f, "ico {}", sizes),
-      IconInfo::SVG => write!(f, "svg"),
+      IconInfo::SVG { size } => {
+        write!(
+          f,
+          "svg{}",
+          if let Some(size) = size {
+            format!(" {}", size)
+          } else {
+            "".to_string()
+          }
+        )
+      }
     }
   }
 }
@@ -187,15 +222,20 @@ impl Display for IconInfo {
 impl Ord for IconInfo {
   fn cmp(&self, other: &Self) -> Ordering {
     match (self, other) {
-      (IconInfo::SVG, IconInfo::SVG) => Ordering::Equal,
-      (IconInfo::SVG, _) => Ordering::Less,
-      (_, IconInfo::SVG) => Ordering::Greater,
+      (IconInfo::SVG { size }, IconInfo::SVG { size: other_size }) => match (size, other_size) {
+        (Some(_), None) => Ordering::Less,
+        (None, Some(_)) => Ordering::Greater,
+        (Some(size), Some(other_size)) => size.cmp(other_size),
+        (None, None) => Ordering::Equal,
+      },
+      (IconInfo::SVG { .. }, _) => Ordering::Less,
+      (_, IconInfo::SVG { .. }) => Ordering::Greater,
 
       _ => {
-        let this_size = self.size().unwrap();
+        let size = self.size().unwrap();
         let other_size = other.size().unwrap();
 
-        this_size.cmp(other_size).then_with(|| match (self, other) {
+        size.cmp(other_size).then_with(|| match (self, other) {
           (IconInfo::PNG { .. }, IconInfo::PNG { .. }) => Ordering::Equal,
           (IconInfo::PNG { .. }, _) => Ordering::Less,
           (_, IconInfo::PNG { .. }) => Ordering::Greater,

src/icon_size/mod.rs

@@ -3,12 +3,14 @@ mod ico;
 mod icon_sizes;
 mod jpeg;
 mod png;
+mod svg;
 
 pub use gif::*;
 pub use ico::*;
 pub use icon_sizes::*;
 pub use jpeg::*;
 pub use png::*;
+pub use svg::*;
 
 use serde::{de, Deserialize, Deserializer, Serialize, Serializer};
 use std::{

src/icon_size/svg.rs

@@ -0,0 +1,84 @@
+use super::IconSize;
+use futures::prelude::*;
+use lol_html::{element, errors::RewritingError, HtmlRewriter, Settings};
+use std::{
+  error::Error,
+  fmt::{self, Display},
+};
+
+#[derive(Debug)]
+struct SizeResult(Option<IconSize>);
+
+impl Display for SizeResult {
+  fn fmt(&self, _: &mut fmt::Formatter<'_>) -> fmt::Result {
+    Ok(())
+  }
+}
+
+impl Error for SizeResult {}
+
+fn parse_size<S: ToString>(size: S) -> Option<u32> {
+  size
+    .to_string()
+    .parse::<f64>()
+    .ok()
+    .map(|size| size.round() as u32)
+}
+
+pub async fn get_svg_size<R: AsyncRead + Unpin>(
+  first_bytes: &[u8; 2],
+  reader: &mut R,
+) -> Result<Option<IconSize>, Box<dyn Error>> {
+  let mut rewriter = HtmlRewriter::new(
+    Settings {
+      element_content_handlers: vec![
+        // Rewrite insecure hyperlinks
+        element!("svg", |el| {
+          let viewbox = el.get_attribute("viewbox");
+
+          let width = el.get_attribute("width").and_then(parse_size);
+          let height = el.get_attribute("height").and_then(parse_size);
+
+          Err(Box::new(SizeResult(
+            if let (Some(width), Some(height)) = (width, height) {
+              Some(IconSize::new(width, height))
+            } else if let Some(viewbox) = viewbox {
+              regex!(r"^\d+\s+\d+\s+(\d+\.?[\d]?)\s+(\d+\.?[\d]?)")
+                .captures(&viewbox)
+                .map(|captures| {
+                  let width = parse_size(captures.get(1).unwrap().as_str()).unwrap();
+                  let height = parse_size(captures.get(2).unwrap().as_str()).unwrap();
+                  IconSize::new(width, height)
+                })
+            } else {
+              None
+            },
+          )))
+        }),
+      ],
+      ..Settings::default()
+    },
+    |_: &[u8]| {},
+  );
+
+  rewriter.write(first_bytes)?;
+
+  let mut buffer = [0; 100];
+
+  loop {
+    let n = reader.read(&mut buffer).await?;
+    if n == 0 {
+      return Err("invalid svg".into());
+    }
+
+    match rewriter.write(&buffer[..n]) {
+      Err(RewritingError::ContentHandlerError(result)) => {
+        let result = result.downcast::<SizeResult>().unwrap();
+
+        return Ok(result.0);
+      }
+
+      result => result?,
+    }
+  }
+}