Implement Get::html() for all platforms (#163)

* implement get html operation

Signed-off-by: Gae24 <96017547+Gae24@users.noreply.github.com>

examples/set_get_html.rs

@@ -15,4 +15,7 @@ consectetur adipiscing elit."#;
 
 	ctx.set_html(html, Some(alt_text)).unwrap();
 	thread::sleep(Duration::from_secs(5));
+
+	let success = ctx.get().html().unwrap() == html;
+	println!("Set and Get html operations were successful: {success}");
 }

src/lib.rs

@@ -192,6 +192,11 @@ impl Get<'_> {
 	pub fn image(self) -> Result<ImageData<'static>, Error> {
 		self.platform.image()
 	}
+
+	/// Completes the "get" operation by fetching HTML from the clipboard.
+	pub fn html(self) -> Result<String, Error> {
+		self.platform.html()
+	}
 }
 
 /// A builder for an operation that sets a value to the clipboard.
@@ -322,6 +327,24 @@ mod tests {
 			ctx.set_html(html, Some(alt_text)).unwrap();
 			assert_eq!(ctx.get_text().unwrap(), alt_text);
 		}
+		{
+			let mut ctx = Clipboard::new().unwrap();
+
+			let html = "<b>hello</b> <i>world</i>!";
+
+			ctx.set().html(html, None).unwrap();
+
+			if cfg!(target_os = "macos") {
+				// Copying HTML on macOS adds wrapper content to work around
+				// historical platform bugs. We control this wrapper, so we are
+				// able to check that the full user data still appears and at what
+				// position in the final copy contents.
+				let content = ctx.get().html().unwrap();
+				assert!(content.ends_with(&format!("{html}</body></html>")));
+			} else {
+				assert_eq!(ctx.get().html().unwrap(), html);
+			}
+		}
 		#[cfg(feature = "image-data")]
 		{
 			let mut ctx = Clipboard::new().unwrap();

src/platform/linux/mod.rs

@@ -122,6 +122,14 @@ impl<'clipboard> Get<'clipboard> {
 			Clipboard::WlDataControl(clipboard) => clipboard.get_image(self.selection),
 		}
 	}
+
+	pub(crate) fn html(self) -> Result<String, Error> {
+		match self.clipboard {
+			Clipboard::X11(clipboard) => clipboard.get_html(self.selection),
+			#[cfg(feature = "wayland-data-control")]
+			Clipboard::WlDataControl(clipboard) => clipboard.get_html(self.selection),
+		}
+	}
 }
 
 /// Linux-specific extensions to the [`Get`](super::Get) builder.

src/platform/linux/wayland.rs

@@ -53,10 +53,12 @@ impl Clipboard {
 		Ok(Self {})
 	}
 
-	pub(crate) fn get_text(&mut self, selection: LinuxClipboardKind) -> Result<String, Error> {
-		use wl_clipboard_rs::paste::MimeType;
-
-		let result = get_contents(selection.try_into()?, Seat::Unspecified, MimeType::Text);
+	fn string_for_mime(
+		&mut self,
+		selection: LinuxClipboardKind,
+		mime: paste::MimeType,
+	) -> Result<String, Error> {
+		let result = get_contents(selection.try_into()?, Seat::Unspecified, mime);
 		match result {
 			Ok((mut pipe, _)) => {
 				let mut contents = vec![];
@@ -74,6 +76,10 @@ impl Clipboard {
 		}
 	}
 
+	pub(crate) fn get_text(&mut self, selection: LinuxClipboardKind) -> Result<String, Error> {
+		self.string_for_mime(selection, paste::MimeType::Text)
+	}
+
 	pub(crate) fn set_text(
 		&self,
 		text: Cow<'_, str>,
@@ -91,6 +97,10 @@ impl Clipboard {
 		Ok(())
 	}
 
+	pub(crate) fn get_html(&mut self, selection: LinuxClipboardKind) -> Result<String, Error> {
+		self.string_for_mime(selection, paste::MimeType::Specific("text/html"))
+	}
+
 	pub(crate) fn set_html(
 		&self,
 		html: Cow<'_, str>,

src/platform/linux/x11.rs

@@ -885,6 +885,12 @@ impl Clipboard {
 		self.inner.write(data, selection, wait)
 	}
 
+	pub(crate) fn get_html(&self, selection: LinuxClipboardKind) -> Result<String> {
+		let formats = [self.inner.atoms.HTML];
+		let result = self.inner.read(&formats, selection)?;
+		String::from_utf8(result.bytes).map_err(|_| Error::ConversionFailure)
+	}
+
 	pub(crate) fn set_html(
 		&self,
 		html: Cow<'_, str>,

src/platform/osx.rs

@@ -121,6 +121,27 @@ impl Clipboard {
 		unsafe { self.pasteboard.clearContents() };
 	}
 
+	fn string_from_type(&self, type_: &'static NSString) -> Result<String, Error> {
+		// XXX: There does not appear to be an alternative for obtaining text without the need for
+		// autorelease behavior.
+		autoreleasepool(|_| {
+			// XXX: We explicitly use `pasteboardItems` and not `stringForType` since the latter will concat
+			// multiple strings, if present, into one and return it instead of reading just the first which is `arboard`'s
+			// historical behavior.
+			let contents = unsafe { self.pasteboard.pasteboardItems() }.ok_or_else(|| {
+				Error::Unknown { description: String::from("NSPasteboard#pasteboardItems errored") }
+			})?;
+
+			for item in contents {
+				if let Some(string) = unsafe { item.stringForType(type_) } {
+					return Ok(string.to_string());
+				}
+			}
+
+			Err(Error::ContentNotAvailable)
+		})
+	}
+
 	// fn get_binary_contents(&mut self) -> Result<Option<ClipboardContent>, Box<dyn std::error::Error>> {
 	// 	let string_class: Id<NSObject> = {
 	// 		let cls: Id<Class> = unsafe { Id::from_ptr(class("NSString")) };
@@ -182,27 +203,11 @@ impl<'clipboard> Get<'clipboard> {
 	}
 
 	pub(crate) fn text(self) -> Result<String, Error> {
-		// XXX: There does not appear to be an alternative for obtaining text without the need for
-		// autorelease behavior.
-		autoreleasepool(|_| {
-			// XXX: We explicitly use `pasteboardItems` and not `stringForType` since the latter will concat
-			// multiple strings, if present, into one and return it instead of reading just the first which is `arboard`'s
-			// historical behavior.
-			let contents =
-				unsafe { self.clipboard.pasteboard.pasteboardItems() }.ok_or_else(|| {
-					Error::Unknown {
-						description: String::from("NSPasteboard#pasteboardItems errored"),
-					}
-				})?;
+		unsafe { self.clipboard.string_from_type(NSPasteboardTypeString) }
+	}
 
-			for item in contents {
-				if let Some(string) = unsafe { item.stringForType(NSPasteboardTypeString) } {
-					return Ok(string.to_string());
-				}
-			}
-
-			Err(Error::ContentNotAvailable)
-		})
+	pub(crate) fn html(self) -> Result<String, Error> {
+		unsafe { self.clipboard.string_from_type(NSPasteboardTypeHTML) }
 	}
 
 	#[cfg(feature = "image-data")]

src/platform/windows.rs

@@ -567,6 +567,19 @@ impl<'clipboard> Get<'clipboard> {
 		String::from_utf16(&out[..bytes_read]).map_err(|_| Error::ConversionFailure)
 	}
 
+	pub(crate) fn html(self) -> Result<String, Error> {
+		let _clipboard_assertion = self.clipboard?;
+
+		let format = clipboard_win::register_format("HTML Format")
+			.ok_or_else(|| Error::unknown("unable to register HTML format"))?;
+
+		let mut out: Vec<u8> = Vec::new();
+		clipboard_win::raw::get_html(format.get(), &mut out)
+			.map_err(|_| Error::unknown("failed to read clipboard string"))?;
+
+		String::from_utf8(out).map_err(|_| Error::ConversionFailure)
+	}
+
 	#[cfg(feature = "image-data")]
 	pub(crate) fn image(self) -> Result<ImageData<'static>, Error> {
 		const FORMAT: u32 = clipboard_win::formats::CF_DIBV5;