Progressing #77

ajrcarey · Mar 17, 2023 · 524019d · 524019d
1 parent e30eec2
commit 524019d
Show file tree

Hide file tree

Showing 3 changed files with 66 additions and 46 deletions.
diff --git a/README.md b/README.md
@@ -236,7 +236,7 @@ a PDF document longer than just a few pages will result in an unrecoverable out
 The WASM builds of Pdfium at <https://github.com/paulocoutinhox/pdfium-lib/releases> are recommended
 as they do not have this problem.
 
-## Multithreading
+## Multi-threading
 
 Pdfium makes no guarantees about thread safety and should be assumed _not_ to be thread safe.
 The Pdfium authors specifically recommend that parallel processing, not multi-threading,
@@ -335,7 +335,7 @@ functions related to page rendering were prioritised. By 1.0, the functionality
 `FPDF_*` functions exported by all Pdfium modules will be available, with the exception of certain
 functions specific to interactive scripting, user interaction, and printing.
 
-* Releases numbered 0.4.x added support for all page rendering Pdfium functions to `pdfium-render`.
+* Releases numbered 0.4.x added support for basic page rendering Pdfium functions to `pdfium-render`.
 * Releases numbered 0.5.x-0.6.x added support for most read-only Pdfium functions to `pdfium-render`.
 * Releases numbered 0.7.x added support for most Pdfium page object creation and editing functions to `pdfium-render`. 
 * Releases numbered 0.8.x aim to progressively add support for all remaining Pdfium editing functions to `pdfium-render`.
@@ -356,7 +356,8 @@ at <https://github.com/ajrcarey/pdfium-render/issues>.
 
 * 0.8.0: removes the ability to acquire an owned `PdfPages` instance from `PdfDocument::pages()`
   as per <https://github.com/ajrcarey/pdfium-render/issues/47>; adds new `PdfDocument::pages_mut()`
-  function to match reworked `PdfDocument::pages()` function.
+  function to match reworked `PdfDocument::pages()` function; fixes a bug in the WASM implementation
+  of `FPDFText_GetBoundedText` as detailed at <https://github.com/ajrcarey/pdfium-render/issues/77>.
 * 0.7.34: replaces functions in `PdfPageLinks` using linear traversal with binary search traversal;
   adds new `PdfFormField` enum; renames `PdfPageObjectFormFragment` to `PdfPageXObjectFormObject`
   to disambiguate it from `PdfForm` and `PdfFormField`; adds `PdfPageAnnotationCommon::as_form_field()`

diff --git a/examples/wasm.rs b/examples/wasm.rs
@@ -30,73 +30,90 @@ pub async fn log_page_metrics_to_console(url: String) {
     log::info!("PDF file version: {:#?}", document.version());
 
     log::info!("PDF metadata tags:");
+
     document
         .metadata()
         .iter()
         .enumerate()
         .for_each(|(index, tag)| log::info!("{}: {:#?} = {}", index, tag.tag_type(), tag.value()));
 
+    let pages = document.pages();
+
     match document.form() {
-        Some(form) => log::info!(
-            "PDF contains an embedded form of type {:#?}",
-            form.form_type()
-        ),
+        Some(form) => {
+            log::info!(
+                "PDF contains an embedded form of type {:#?}",
+                form.form_type()
+            );
+
+            for (key, value) in form.field_values(&pages).iter() {
+                log::info!("{:?} => {:?}", key, value);
+            }
+        }
         None => log::info!("PDF does not contain an embedded form"),
     };
 
     // Report labels, boundaries, and metrics for each page to the console.
 
-    document
-        .pages()
-        .iter()
-        .enumerate()
-        .for_each(|(page_index, page)| {
-            if let Some(label) = page.label() {
-                log::info!("Page {} has a label: {}", page_index, label);
-            }
+    pages.iter().enumerate().for_each(|(page_index, page)| {
+        if let Some(label) = page.label() {
+            log::info!("Page {} has a label: {}", page_index, label);
+        }
+
+        log::info!(
+            "Page {} width: {}, height: {}",
+            page_index,
+            page.width().value,
+            page.height().value
+        );
 
+        for boundary in page.boundaries().iter() {
             log::info!(
-                "Page {} width: {}, height: {}",
+                "Page {} has defined {:#?} box ({}, {}) - ({}, {})",
                 page_index,
-                page.width().value,
-                page.height().value
+                boundary.box_type,
+                boundary.bounds.left.value,
+                boundary.bounds.top.value,
+                boundary.bounds.right.value,
+                boundary.bounds.bottom.value,
             );
+        }
 
-            for boundary in page.boundaries().iter() {
-                log::info!(
-                    "Page {} has defined {:#?} box ({}, {}) - ({}, {})",
-                    page_index,
-                    boundary.box_type,
-                    boundary.bounds.left.value,
-                    boundary.bounds.top.value,
-                    boundary.bounds.right.value,
-                    boundary.bounds.bottom.value,
-                );
-            }
+        log::info!(
+            "Page {} has paper size {:#?}",
+            page_index,
+            page.paper_size()
+        );
 
+        for (link_index, link) in page.links().iter().enumerate() {
             log::info!(
-                "Page {} has paper size {:#?}",
+                "Page {} link {} has action of type {:?}",
                 page_index,
-                page.paper_size()
+                link_index,
+                link.action().map(|action| action.action_type())
             );
 
-            for (link_index, link) in page.links().iter().enumerate() {
-                log::info!(
-                    "Page {} link {} has action of type {:?}",
-                    page_index,
-                    link_index,
-                    link.action().map(|action| action.action_type())
-                );
+            // For links that have URI actions, output the destination URI.
 
-                // For links that have URI actions, output the destination URI.
-
-                if let Some(action) = link.action() {
-                    if let Some(uri_action) = action.as_uri_action() {
-                        log::info!("Link URI destination: {:#?}", uri_action.uri())
-                    }
+            if let Some(action) = link.action() {
+                if let Some(uri_action) = action.as_uri_action() {
+                    log::info!("Link URI destination: {:#?}", uri_action.uri())
                 }
             }
-        });
+        }
+
+        let text = page.text().unwrap();
+
+        for (annotation_index, annotation) in page.annotations().iter().enumerate() {
+            log::info!(
+                "Page {} annotation {} has text: {:?}, bounds: {:?}",
+                page_index,
+                annotation_index,
+                text.for_annotation(&annotation),
+                annotation.bounds()
+            );
+        }
+    });
 }
 
 /// Downloads the given url, opens it as a PDF document, then returns the ImageData for

diff --git a/src/wasm.rs b/src/wasm.rs
@@ -7534,7 +7534,9 @@ impl PdfiumLibraryBindings for WasmPdfiumBindings {
             .unwrap() as usize;
 
         if result > 0 && result <= buffer_length {
-            state.copy_struct_from_pdfium(buffer_ptr, result, buffer);
+            // The return result is the number of _characters_ returned, _not_ the number of bytes.
+
+            state.copy_struct_from_pdfium(buffer_ptr, result * size_of::<c_ushort>(), buffer);
         }
 
         state.free(buffer_ptr);