Add initial OCR support in document_writer

.github/workflows/CI.yml

@@ -38,7 +38,7 @@ jobs:
           submodules: "recursive"
           fetch-depth: 500
       - uses: dtolnay/rust-toolchain@stable
-      - run: sudo apt-get -y install libfontconfig1-dev
+      - run: sudo apt update && sudo apt-get -y install libfontconfig1-dev libc++abi-dev clang gcc g++
         if: matrix.os == 'ubuntu-latest'
       - name: Install LLVM
         if: matrix.os == 'windows-2019'
@@ -74,7 +74,7 @@ jobs:
       - uses: msys2/setup-msys2@v2
         with:
           msystem: ${{matrix.sys}}
-          install: mingw-w64-${{matrix.env}}-rust mingw-w64-${{matrix.env}}-clang base base-devel unzip git
+          install: mingw-w64-${{matrix.env}}-rust mingw-w64-${{matrix.env}}-clang base base-devel unzip git python3
 
       - uses: actions/checkout@v3
         with:
@@ -95,7 +95,7 @@ jobs:
       - uses: dtolnay/rust-toolchain@nightly
         with:
           components: rust-src
-      - run: sudo apt-get -y install libfontconfig1-dev
+      - run: sudo apt update && sudo apt-get -y install libfontconfig1-dev libc++abi-dev clang gcc g++
       - name: cargo test --features serde
         run: |
           cargo test -Zbuild-std --target x86_64-unknown-linux-gnu --features serde
@@ -113,7 +113,7 @@ jobs:
           fetch-depth: 500
       - uses: dtolnay/rust-toolchain@stable
       - uses: taiki-e/install-action@valgrind
-      - run: sudo apt-get -y install libfontconfig1-dev
+      - run: sudo apt update && sudo apt-get -y install libfontconfig1-dev libc++abi-dev clang gcc g++
       - run: cargo test --features serde
         env:
           CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_RUNNER: "valgrind --error-exitcode=1 --track-origins=yes"

mupdf-sys/build.rs

@@ -48,7 +48,7 @@ const CPU_FLAGS: &[(&str, &str, &str, Option<&str>)] = &[
     ("avx", "-mavx", "HAVE_AVX", None),
     ("avx2", "-mavx2", "HAVE_AVX2", None),
     ("fma", "-mfma", "HAVE_FMA", None),
-    ("neon", "-mfpu=neon", "HAVE_NEON", Some("ARCH_HAS_NEON")),
+    //("neon", "-mfpu=neon", "HAVE_NEON", Some("ARCH_HAS_NEON")),
 ];
 
 #[cfg(not(target_env = "msvc"))]
@@ -99,14 +99,15 @@ fn build_libmupdf() {
         "libs".to_owned(),
         format!("build={}", profile),
         format!("OUT={}", &build_dir_str),
-        #[cfg(not(feature = "tesseract"))]
-        "USE_TESSERACT=no".to_owned(),
+        #[cfg(feature = "tesseract")]
+        "USE_TESSERACT=yes".to_owned(),
         #[cfg(not(feature = "libarchive"))]
         "USE_LIBARCHIVE=no".to_owned(),
         #[cfg(not(feature = "zxingcpp"))]
         "USE_ZXINGCPP=no".to_owned(),
         #[cfg(feature = "sys-lib")]
         "USE_SYSTEM_LIBS=yes".to_owned(),
+        "USE_ARGUMENT_FILE=yes".to_owned(),
         "HAVE_X11=no".to_owned(),
         "HAVE_GLUT=no".to_owned(),
         "HAVE_CURL=no".to_owned(),
@@ -513,6 +514,16 @@ fn main() {
     println!("cargo:rerun-if-changed=wrapper.h");
     println!("cargo:rerun-if-changed=wrapper.c");
 
+    if let Ok(ref target_os) = env::var("CARGO_CFG_TARGET_OS") {
+        if target_os != "windows" {
+            if target_os == "macos" {
+                println!("cargo:rustc-link-lib=c++");
+            } else {
+                println!("cargo:rustc-link-lib=stdc++");
+            }
+        }
+    }
+
     build_libmupdf();
 
     let mut build = cc::Build::new();

mupdf-sys/wrapper.c

@@ -3227,6 +3227,20 @@ fz_document_writer *mupdf_new_document_writer(fz_context *ctx, const char *filen
     return writer;
 }
 
+fz_document_writer *mupdf_new_pdfocr_writer(fz_context *ctx, const char *path, const char *options, mupdf_error_t **errptr)
+{
+    fz_document_writer *writer = NULL;
+    fz_try(ctx)
+    {
+        writer = fz_new_pdfocr_writer(ctx, path, options);
+    }
+    fz_catch(ctx)
+    {
+        mupdf_save_error(ctx, errptr);
+    }
+    return writer;
+}
+
 fz_device *mupdf_document_writer_begin_page(fz_context *ctx, fz_document_writer *writer, fz_rect mediabox, mupdf_error_t **errptr)
 {
     fz_device *device = NULL;

src/document_writer.rs

@@ -26,6 +26,21 @@ impl DocumentWriter {
         .map(|inner| Self { inner })
     }
 
+    #[cfg(feature = "tesseract")]
+    pub fn new_pdfocr_writer(path: &str, options: &str) -> Result<Self, Error> {
+        let c_path = CString::new(path)?;
+        let c_options = CString::new(options)?;
+
+        unsafe {
+            ffi_try!(mupdf_new_pdfocr_writer(
+                context(),
+                c_path.as_ptr(),
+                c_options.as_ptr()
+            ))
+        }
+        .map(|inner| Self { inner })
+    }
+
     pub fn begin_page(&mut self, media_box: Rect) -> Result<Device, Error> {
         unsafe {
             ffi_try!(mupdf_document_writer_begin_page(