Update PDFium binaries to version 4955 and add support for UTF-8 char…

…acters in document filenames. (#62) See original PDFium commit here https://pdfium.googlesource.com/pdfium/+/98e0feb735a547c51f31b1da6aa9004542ef728c Co-authored-by: Dmitry Zhelnin <[email protected]>
GowenGit · Apr 24, 2022 · 476cc6b · 476cc6b
1 parent 9093dab
commit 476cc6b
Show file tree

Hide file tree

Showing 11 changed files with 37 additions and 18 deletions.
diff --git a/README.md b/README.md
@@ -9,7 +9,7 @@
 
 **docnet** aims to be a fast PDF editing and data extraction library. It is a `.NET Standard 2.0` wrapper for `PDFium C++` library that is used by `chromium`.
 
-PDFium version: `4722`
+PDFium version: `4955`
 
 Supported platforms:
 

diff --git a/src/Docnet.Core/Bindings/PdfiumWrapper.cs b/src/Docnet.Core/Bindings/PdfiumWrapper.cs
@@ -21,6 +21,7 @@
 using System.Collections.Concurrent;
 using System.Runtime.InteropServices;
 using System.Security;
+using System.Text;
 
 namespace Docnet.Core.Bindings
 {
@@ -1745,7 +1746,7 @@ public partial struct __Internal
             [DllImport("pdfium", CallingConvention = CallingConvention.Cdecl,
                 EntryPoint = "FPDF_LoadDocument")]
             internal static extern IntPtr FPDF_LoadDocument(
-                [MarshalAs(UnmanagedType.LPStr)] string file_path, [MarshalAs(UnmanagedType.LPStr)] string password);
+                IntPtr file_path, [MarshalAs(UnmanagedType.LPStr)] string password);
 
             [SuppressUnmanagedCodeSecurity]
             [DllImport("pdfium", CallingConvention = CallingConvention.Cdecl,
@@ -1975,16 +1976,33 @@ public static int FPDF_SetPrintMode(int mode)
             return __ret;
         }
 
+        private static IntPtr NativeUtf8FromString(string managedString) {
+            int len = Encoding.UTF8.GetByteCount(managedString);
+            byte[] buffer = new byte[len + 1];
+            Encoding.UTF8.GetBytes(managedString, 0, managedString.Length, buffer, 0);
+            IntPtr nativeUtf8 = Marshal.AllocHGlobal(buffer.Length);
+            Marshal.Copy(buffer, 0, nativeUtf8, buffer.Length);
+            return nativeUtf8;
+        }
+
         public static FpdfDocumentT FPDF_LoadDocument(string file_path, string password)
         {
-            var __ret = __Internal.FPDF_LoadDocument(file_path, password);
-            FpdfDocumentT __result0;
-            if (__ret == IntPtr.Zero) __result0 = null;
-            else if (FpdfDocumentT.NativeToManagedMap.ContainsKey(__ret))
-                __result0 = (FpdfDocumentT)FpdfDocumentT
-                    .NativeToManagedMap[__ret];
-            else __result0 = FpdfDocumentT.__CreateInstance(__ret);
-            return __result0;
+            var ptr = NativeUtf8FromString(file_path);
+            try
+            {
+                var __ret = __Internal.FPDF_LoadDocument(ptr, password);
+                FpdfDocumentT __result0;
+                if (__ret == IntPtr.Zero) __result0 = null;
+                else if (FpdfDocumentT.NativeToManagedMap.ContainsKey(__ret))
+                    __result0 = (FpdfDocumentT)FpdfDocumentT
+                        .NativeToManagedMap[__ret];
+                else __result0 = FpdfDocumentT.__CreateInstance(__ret);
+                return __result0;
+            }
+            finally
+            {
+                Marshal.FreeHGlobal(ptr);
+            }
         }
 
         public static FpdfDocumentT FPDF_LoadMemDocument(IntPtr data_buf,

diff --git a/src/Docnet.Core/runtimes/linux-arm/native/pdfium.so b/src/Docnet.Core/runtimes/linux-arm/native/pdfium.so
diff --git a/src/Docnet.Core/runtimes/linux-arm64/native/pdfium.so b/src/Docnet.Core/runtimes/linux-arm64/native/pdfium.so
diff --git a/src/Docnet.Core/runtimes/linux/native/pdfium.so b/src/Docnet.Core/runtimes/linux/native/pdfium.so
diff --git a/src/Docnet.Core/runtimes/osx/native/pdfium.dylib b/src/Docnet.Core/runtimes/osx/native/pdfium.dylib
diff --git a/src/Docnet.Core/runtimes/win-x64/native/pdfium.dll b/src/Docnet.Core/runtimes/win-x64/native/pdfium.dll
diff --git a/src/Docnet.Core/runtimes/win-x86/native/pdfium.dll b/src/Docnet.Core/runtimes/win-x86/native/pdfium.dll
diff --git a/src/Docnet.Tests.Integration/DocReaderTests.cs b/src/Docnet.Tests.Integration/DocReaderTests.cs
@@ -63,6 +63,7 @@ public void GetDocReader_WhenCalledWithInvalidDimensions_ShouldThrow(Input type,
         [InlineData(Input.FromBytes, "Docs/simple_3.pdf", null, 2)]
         [InlineData(Input.FromFile, "Docs/protected_0.pdf", "password", 3)]
         [InlineData(Input.FromBytes, "Docs/protected_0.pdf", "password", 3)]
+        [InlineData(Input.FromFile, "Docs/fancy_doc_∮.pdf", null, 2)]
         public void GetPageCount_WhenCalled_ShouldReturnCorrectResults(Input type, string filePath, string password, int expectedCount)
         {
             using (var reader = _fixture.GetDocReader(type, filePath, password, 10, 10))

diff --git a/src/Docnet.Tests.Integration/Docs/fancy_doc_∮.pdf b/src/Docnet.Tests.Integration/Docs/fancy_doc_∮.pdf
diff --git a/tools/get_pdfium.sh b/tools/get_pdfium.sh
@@ -1,11 +1,11 @@
 #!/bin/bash
 
-wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4722/pdfium-linux-x64.tgz
-wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4722/pdfium-linux-arm.tgz
-wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4722/pdfium-linux-arm64.tgz
-wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4722/pdfium-win-x64.tgz
-wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4722/pdfium-win-x86.tgz
-wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4722/pdfium-mac-x64.tgz
+wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4955/pdfium-linux-x64.tgz
+wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4955/pdfium-linux-arm.tgz
+wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4955/pdfium-linux-arm64.tgz
+wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4955/pdfium-win-x64.tgz
+wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4955/pdfium-win-x86.tgz
+wget https://github.com/bblanchon/pdfium-binaries/releases/download/chromium%2F4955/pdfium-mac-x64.tgz
 
 mkdir linux
 mkdir linux-arm
@@ -40,10 +40,10 @@ cp linux-arm64/LICENSE ../src/Docnet.Core/runtimes/linux-arm64/native/LICENSE
 cp osx/lib/libpdfium.dylib ../src/Docnet.Core/runtimes/osx/native/pdfium.dylib
 cp osx/LICENSE ../src/Docnet.Core/runtimes/osx/native/LICENSE
 
-cp windows/x64/bin/pdfium.dll ../src/Docnet.Core/runtimes/win-x64/native/pdfium.dll
+cp windows/bin/pdfium.dll ../src/Docnet.Core/runtimes/win-x64/native/pdfium.dll
 cp windows/LICENSE ../src/Docnet.Core/runtimes/win-x64/native/LICENSE
 
-cp windowsx86/x86/bin/pdfium.dll ../src/Docnet.Core/runtimes/win-x86/native/pdfium.dll
+cp windowsx86/bin/pdfium.dll ../src/Docnet.Core/runtimes/win-x86/native/pdfium.dll
 cp windowsx86/LICENSE ../src/Docnet.Core/runtimes/win-x86/native/LICENSE
 
 rm pdfium-linux-x64.tgz pdfium-linux-arm.tgz pdfium-linux-arm64.tgz pdfium-win-x64.tgz pdfium-win-x86.tgz pdfium-mac-x64.tgz