Progressing #145

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "pdfium-render"
-version = "0.8.22"
+version = "0.8.23"
 edition = "2018"
 publish = true
 description = "A high-level idiomatic Rust wrapper around Pdfium, the C++ PDF library used by the Google Chromium project."
@@ -36,25 +36,36 @@ console_error_panic_hook = "0"
 js-sys = "0"
 wasm-bindgen = { version = "0", features = ["enable-interning"] }
 wasm-bindgen-futures = { version = "0" }
-web-sys = { version = "0", features = ["TextDecoder", "ImageData", "Window", "Response", "Blob"] }
+web-sys = { version = "0", features = [
+    "TextDecoder",
+    "ImageData",
+    "Window",
+    "Response",
+    "Blob",
+] }
 
 [target.'cfg(not(target_arch = "wasm32"))'.dependencies]
-libloading = "0"
+libloading = "0.8.1"
 
 [build-dependencies]
 bindgen = { version = "0" }
 
 [dev-dependencies]
 # Dependencies specific to examples. Dependencies for the WASM example in examples/wasm.rs
 # are specified separately in examples/Cargo.toml.
-rand = "0" # Used by examples/create.rs, examples/matrix.rs
+rand = "0"  # Used by examples/create.rs, examples/matrix.rs
 rayon = "1" # Used by examples/thread_safe.rs
 
 [features]
-default = ["image", "thread_safe"] # The image feature is implicit because the image dependency is optional.
+default = [
+    "image",
+    "thread_safe",
+] # The image feature is implicit because the image dependency is optional.
 bindings = []
 static = []
 thread_safe = []
 "libstdc++" = ["static"]
 "libc++" = ["static"]
 sync = ["thread_safe"]
+paragraph = []
+flatten = []

src/attachment.rs

@@ -173,9 +173,9 @@ impl<'a> PdfAttachment<'a> {
     /// This function is not available when compiling to WASM. You have several options for
     /// saving attachment data in WASM:
     /// * Use either the [PdfAttachment::save_to_writer()] or the [PdfAttachment::save_to_bytes()] functions,
-    /// both of which are available when compiling to WASM.
+    ///   both of which are available when compiling to WASM.
     /// * Use the [PdfAttachment::save_to_blob()] function to save attachment data directly into a new
-    /// Javascript `Blob` object. This function is only available when compiling to WASM.
+    ///   Javascript `Blob` object. This function is only available when compiling to WASM.
     #[cfg(not(target_arch = "wasm32"))]
     pub fn save_to_file(&self, path: &(impl AsRef<Path> + ?Sized)) -> Result<(), PdfiumError> {
         self.save_to_writer(&mut File::create(path).map_err(PdfiumError::IoError)?)

src/attachments.rs

@@ -158,15 +158,15 @@ impl<'a> PdfAttachments<'a> {
     /// This function is not available when compiling to WASM. You have several options for
     /// loading attachment data in WASM:
     /// * Use the [PdfAttachments::create_attachment_from_fetch()] function to download attachment data
-    /// from a URL using the browser's built-in `fetch()` API. This function is only available when
-    /// compiling to WASM.
+    ///   from a URL using the browser's built-in `fetch()` API. This function is only available when
+    ///   compiling to WASM.
     /// * Use the [PdfAttachments::create_attachment_from_blob()] function to load attachment data
-    /// from a Javascript `File` or `Blob` object (such as a `File` object returned from an HTML
-    /// `<input type="file">` element). This function is only available when compiling to WASM.
+    ///   from a Javascript `File` or `Blob` object (such as a `File` object returned from an HTML
+    ///   `<input type="file">` element). This function is only available when compiling to WASM.
     /// * Use another method to retrieve the bytes of the target attachment over the network,
-    /// then load those bytes into Pdfium using the [PdfAttachments::create_attachment_from_bytes()] function.
+    ///   then load those bytes into Pdfium using the [PdfAttachments::create_attachment_from_bytes()] function.
     /// * Embed the bytes of the target attachment directly into the compiled WASM module
-    /// using the `include_bytes!()` macro.
+    ///   using the `include_bytes!()` macro.
     #[cfg(not(target_arch = "wasm32"))]
     pub fn create_attachment_from_file(
         &mut self,

src/color_space.rs

@@ -18,20 +18,20 @@ use crate::error::PdfiumError;
 /// share the same general characteristics. Families fall into three broad categories:
 ///
 /// * **Device color spaces** directly specify colors or shades of gray that the output
-/// device is to produce. The precise displayed color is device-specific and is not calibrated.
-/// Color space families in this category include [PdfColorSpace::DeviceGray],
-/// [PdfColorSpace::DeviceRGB], and [PdfColorSpace::DeviceCMYK].
+///   device is to produce. The precise displayed color is device-specific and is not calibrated.
+///   Color space families in this category include [PdfColorSpace::DeviceGray],
+///   [PdfColorSpace::DeviceRGB], and [PdfColorSpace::DeviceCMYK].
 /// * **Calibrated color spaces** are based on international standards for specifying human-visible
-/// colors created by the Commission Internationale de l'Éclairage (International Commission on
-/// Illumination) and the International Color Consortium. The precise displayed color is
-/// device-independent; it does not rely on the characteristics of any particular output device.
-/// Color space families in this category include [PdfColorSpace::CalibratedCIEGray],
-/// [PdfColorSpace::CalibratedCIERGB], [PdfColorSpace::CalibratedCIELab], and
-/// [PdfColorSpace::CalibratedICCProfile].
+///   colors created by the Commission Internationale de l'Éclairage (International Commission on
+///   Illumination) and the International Color Consortium. The precise displayed color is
+///   device-independent; it does not rely on the characteristics of any particular output device.
+///   Color space families in this category include [PdfColorSpace::CalibratedCIEGray],
+///   [PdfColorSpace::CalibratedCIERGB], [PdfColorSpace::CalibratedCIELab], and
+///   [PdfColorSpace::CalibratedICCProfile].
 /// * **Special color spaces** add features or properties to another color space, such as
-/// patterns, color mapping, separations, and high-fidelity and/or multi-tone color.
-/// Color space families in this category include [PdfColorSpace::Pattern],
-/// [PdfColorSpace::Indexed], [PdfColorSpace::Separation], and [PdfColorSpace::DeviceN].
+///   patterns, color mapping, separations, and high-fidelity and/or multi-tone color.
+///   Color space families in this category include [PdfColorSpace::Pattern],
+///   [PdfColorSpace::Indexed], [PdfColorSpace::Separation], and [PdfColorSpace::DeviceN].
 ///
 /// Non-RGB color spaces typically define a transform that enables color values in the color
 /// space to be converted to an RGB approximation for on-screen display.

src/document.rs

@@ -137,7 +137,7 @@ impl PdfDocumentVersion {
 /// * [PdfDocument::pages()], an immutable collection of all the [PdfPages] in the document.
 /// * [PdfDocument::pages_mut()], a mutable collection of all the [PdfPages] in the document.
 /// * [PdfDocument::permissions()], settings relating to security handlers and document permissions
-/// for the document.
+///   for the document.
 /// * [PdfDocument::signatures()], an immutable collection of all the [PdfSignatures] in the document.
 pub struct PdfDocument<'a> {
     handle: FPDF_DOCUMENT,
@@ -344,9 +344,9 @@ impl<'a> PdfDocument<'a> {
     /// This function is not available when compiling to WASM. You have several options for
     /// saving your PDF document data in WASM:
     /// * Use either the [PdfDocument::save_to_writer()] or the [PdfDocument::save_to_bytes()] functions,
-    /// both of which are available when compiling to WASM.
+    ///   both of which are available when compiling to WASM.
     /// * Use the [PdfDocument::save_to_blob()] function to save document data directly into a new
-    /// Javascript `Blob` object. This function is only available when compiling to WASM.
+    ///   Javascript `Blob` object. This function is only available when compiling to WASM.
     #[cfg(not(target_arch = "wasm32"))]
     pub fn save_to_file(&self, path: &(impl AsRef<Path> + ?Sized)) -> Result<(), PdfiumError> {
         self.save_to_writer(&mut File::create(path).map_err(PdfiumError::IoError)?)

src/flatten.rs

@@ -0,0 +1,7 @@
+use crate::bindgen::FPDF_PAGE;
+use crate::error::PdfiumError;
+
+#[allow(dead_code)] // During development of feature
+fn flatten(_page_handle: FPDF_PAGE) -> Result<(), PdfiumError> {
+    unimplemented!()
+}

src/fonts.rs

@@ -263,17 +263,17 @@ impl<'a> PdfFonts<'a> {
     /// This function is not available when compiling to WASM. You have several options for
     /// loading font data in WASM:
     /// * Use the [PdfFont::load_type1_from_fetch()] function to download font data from a
-    /// URL using the browser's built-in `fetch()` API. This function is only available when
-    /// compiling to WASM.
+    ///   URL using the browser's built-in `fetch()` API. This function is only available when
+    ///   compiling to WASM.
     /// * Use the [PdfFont::load_type1_from_blob()] function to load font data from a
-    /// Javascript File or Blob object (such as a File object returned from an HTML
-    /// `<input type="file">` element). This function is only available when compiling to WASM.
+    ///   Javascript File or Blob object (such as a File object returned from an HTML
+    ///   `<input type="file">` element). This function is only available when compiling to WASM.
     /// * Use the [PdfFont::load_type1_from_reader()] function to load font data from any
-    /// valid Rust reader.
+    ///   valid Rust reader.
     /// * Use another method to retrieve the bytes of the target font over the network,
-    /// then load those bytes into Pdfium using the [PdfFont::new_type1_from_bytes()] function.
+    ///   then load those bytes into Pdfium using the [PdfFont::new_type1_from_bytes()] function.
     /// * Embed the bytes of the desired font directly into the compiled WASM module
-    /// using the `include_bytes!()` macro.
+    ///   using the `include_bytes!()` macro.
     #[cfg(not(target_arch = "wasm32"))]
     pub fn load_type1_from_file(
         &mut self,
@@ -405,17 +405,17 @@ impl<'a> PdfFonts<'a> {
     /// This function is not available when compiling to WASM. You have several options for
     /// loading font data in WASM:
     /// * Use the [PdfFont::load_true_type_from_fetch()] function to download font data from a
-    /// URL using the browser's built-in `fetch()` API. This function is only available when
-    /// compiling to WASM.
+    ///   URL using the browser's built-in `fetch()` API. This function is only available when
+    ///   compiling to WASM.
     /// * Use the [PdfFont::load_true_type_from_blob()] function to load font data from a
-    /// Javascript `File` or `Blob` object (such as a `File` object returned from an HTML
-    /// `<input type="file">` element). This function is only available when compiling to WASM.
+    ///   Javascript `File` or `Blob` object (such as a `File` object returned from an HTML
+    ///   `<input type="file">` element). This function is only available when compiling to WASM.
     /// * Use the [PdfFont::load_true_type_from_reader()] function to load font data from any
-    /// valid Rust reader.
+    ///   valid Rust reader.
     /// * Use another method to retrieve the bytes of the target font over the network,
-    /// then load those bytes into Pdfium using the [PdfFont::new_true_type_from_bytes()] function.
+    ///   then load those bytes into Pdfium using the [PdfFont::new_true_type_from_bytes()] function.
     /// * Embed the bytes of the desired font directly into the compiled WASM module
-    /// using the `include_bytes!()` macro.
+    ///   using the `include_bytes!()` macro.
     #[cfg(not(target_arch = "wasm32"))]
     pub fn load_true_type_from_file(
         &mut self,

src/lib.rs

@@ -31,6 +31,7 @@ pub mod color_space;
 pub mod destination;
 pub mod document;
 pub mod error;
+mod flatten; // Keep internal flatten operation private.
 pub mod font;
 pub mod font_glyph;
 pub mod font_glyphs;
@@ -96,6 +97,7 @@ pub mod page_text_search;
 pub mod page_text_segment;
 pub mod page_text_segments;
 pub mod pages;
+#[cfg(feature = "paragraph")]
 pub mod paragraph;
 pub mod path_segment;
 pub mod path_segments;

src/page.rs

@@ -784,14 +784,14 @@ impl<'a> PdfPage<'a> {
     /// * [PdfPage::translate()]: changes the position of each object on this [PdfPage].
     /// * [PdfPage::scale()]: changes the size of each object on this [PdfPage].
     /// * [PdfPage::flip_horizontally()]: flips each object on this [PdfPage] horizontally around
-    /// the page origin point.
+    ///   the page origin point.
     /// * [PdfPage::flip_vertically()]: flips each object on this [PdfPage] vertically around
-    /// the page origin point.
+    ///   the page origin point.
     /// * [PdfPage::rotate_clockwise_degrees()], [PdfPage::rotate_counter_clockwise_degrees()],
-    /// [PdfPage::rotate_clockwise_radians()], [PdfPage::rotate_counter_clockwise_radians()]:
-    /// rotates each object on this [PdfPage] around its origin.
+    ///   [PdfPage::rotate_clockwise_radians()], [PdfPage::rotate_counter_clockwise_radians()]:
+    ///   rotates each object on this [PdfPage] around its origin.
     /// * [PdfPage::skew_degrees()], [PdfPage::skew_radians()]: skews each object
-    /// on this [PdfPage] relative to its axes.
+    ///   on this [PdfPage] relative to its axes.
     ///
     /// **The order in which transformations are applied is significant.**
     /// For example, the result of rotating _then_ translating an object may be vastly different
@@ -891,6 +891,17 @@ impl<'a> PdfPage<'a> {
     }
 
     /// Flattens all annotations and form fields on this [PdfPage] into the page contents.
+    #[cfg(feature = "flatten")]
+    // Use a custom-written flatten operation, rather than Pdfium's built-in flatten. See:
+    // https://github.com/ajrcarey/pdfium-render/issues/140
+    pub fn flatten(&mut self) -> Result<(), PdfiumError> {
+        flatten_page(self.handle())
+    }
+
+    /// Flattens all annotations and form fields on this [PdfPage] into the page contents.
+    #[cfg(not(feature = "flatten"))]
+    // Use Pdfium's built-in flatten. This has some problems; see:
+    // https://github.com/ajrcarey/pdfium-render/issues/140
     pub fn flatten(&mut self) -> Result<(), PdfiumError> {
         // TODO: AJRC - 28/5/22 - consider allowing the caller to set the FLAT_NORMALDISPLAY or FLAT_PRINT flag.
         let flag = FLAT_PRINT;
@@ -998,6 +1009,7 @@ impl<'a> PdfPage<'a> {
     }
 
     /// Commits any staged but unsaved changes to this [PdfPage] to the underlying [PdfDocument].
+    #[inline]
     pub(crate) fn regenerate_content_immut(&self) -> Result<(), PdfiumError> {
         Self::regenerate_content_immut_for_handle(self.page_handle, self.bindings)
     }

src/page_boundaries.rs

@@ -39,13 +39,13 @@ impl PdfPageBoundaryBox {
 /// The content of a page can be bounded by up to six different boxes:
 ///
 /// * Media box: the full page size, equivalent to the target paper size when the document is printed.
-/// All other page boundaries must fit inside the Media box.
+///   All other page boundaries must fit inside the Media box.
 /// * Art box: the maximum extent of out-of-bleed page art when offset printing.
-/// Typically cropped out when viewing the document on-screen.
+///   Typically cropped out when viewing the document on-screen.
 /// * Bleed box: the maximum extent of outside-trim page bleeds when offset printing.
-/// Typically cropped out when viewing the document on-screen.
+///   Typically cropped out when viewing the document on-screen.
 /// * Trim box: the maximum extent of page trims when offset printing.
-/// Typically cropped out when viewing the document on-screen.
+///   Typically cropped out when viewing the document on-screen.
 /// * Crop box: the maximum extent of user-visible content when viewing the document on-screen.
 /// * Bounding box ("BBox"): the smallest rectangle that can enclose all the content contained in the page.
 ///

src/page_object.rs

@@ -705,14 +705,14 @@ pub trait PdfPageObjectCommon<'a> {
     /// Not all page objects can be successfully copied. The following restrictions apply:
     ///
     /// * For path objects, it is not possible to copy a path object that contains a Bézier path
-    /// segment, because Pdfium does not currently provide any way to retrieve the control points of a
-    /// Bézier curve of an existing path object.
+    ///   segment, because Pdfium does not currently provide any way to retrieve the control points of a
+    ///   Bézier curve of an existing path object.
     /// * For text objects, the font used by the object must be present in the destination document,
-    /// or text rendering behaviour will be unpredictable. While text objects refer to fonts,
-    /// font data is embedded into documents separately from text objects.
+    ///   or text rendering behaviour will be unpredictable. While text objects refer to fonts,
+    ///   font data is embedded into documents separately from text objects.
     /// * For image objects, Pdfium allows iterating over the list of image filters applied
-    /// to an image object, but currently provides no way to set a new object's image filters.
-    /// As a result, it is not possible to copy an image object that has any image filters applied.
+    ///   to an image object, but currently provides no way to set a new object's image filters.
+    ///   As a result, it is not possible to copy an image object that has any image filters applied.
     ///
     /// Pdfium currently allows setting the blend mode for a page object, but provides no way
     /// to retrieve an object's current blend mode. As a result, the blend mode setting of the
@@ -725,14 +725,14 @@ pub trait PdfPageObjectCommon<'a> {
     /// Not all page objects can be successfully copied. The following restrictions apply:
     ///
     /// * For path objects, it is not possible to copy a path object that contains a Bézier path
-    /// segment, because Pdfium does not currently provide any way to retrieve the control points of a
-    /// Bézier curve of an existing path object.
+    ///   segment, because Pdfium does not currently provide any way to retrieve the control points of a
+    ///   Bézier curve of an existing path object.
     /// * For text objects, the font used by the object must be present in the destination document,
-    /// or text rendering behaviour will be unpredictable. While text objects refer to fonts,
-    /// font data is embedded into documents separately from text objects.
+    ///   or text rendering behaviour will be unpredictable. While text objects refer to fonts,
+    ///   font data is embedded into documents separately from text objects.
     /// * For image objects, Pdfium allows iterating over the list of image filters applied
-    /// to an image object, but currently provides no way to set a new object's image filters.
-    /// As a result, it is not possible to copy an image object that has any image filters applied.
+    ///   to an image object, but currently provides no way to set a new object's image filters.
+    ///   As a result, it is not possible to copy an image object that has any image filters applied.
     ///
     /// Pdfium currently allows setting the blend mode for a page object, but provides no way
     /// to retrieve an object's current blend mode. As a result, the blend mode setting of the