cibuildwheel-related work

.github/workflows/cibuildwheel.yaml

@@ -0,0 +1,108 @@
+# SPDX-FileCopyrightText: 2025 wojiushixiaobai <[email protected]>
+# SPDX-FileCopyrightText: 2025 geisserml <[email protected]>
+# SPDX-License-Identifier: Apache-2.0 OR BSD-3-Clause
+
+name: Build with cibuildwheel
+on:
+  workflow_dispatch:
+    inputs:
+      cibw_py_ver:
+        default: 'cp312'
+        type: string
+      existing:
+        default: true
+        type: boolean
+      ibm:
+        default: true
+        type: boolean
+      # emulated:
+      #   default: false
+      #   type: boolean
+
+permissions: {}
+
+jobs:
+
+  build_wheels:
+    name: Build wheel for ${{ matrix.arch }} ${{ matrix.image }} (runs on ${{ matrix.os }})
+    runs-on: ${{ matrix.os }}
+
+    strategy:
+      fail-fast: false
+      matrix:
+        # GitHub doesn't seem to have a proper way to make conditional matrix entries,
+        # so we have to use an `enabled` property and check it on each step (nasty).
+        include:
+          - os: ubuntu-24.04
+            arch: x86_64
+            image: manylinux
+            enabled: ${{ inputs.existing }}
+            needs_emulator: false
+          - os: ubuntu-24.04-arm
+            arch: aarch64
+            image: manylinux
+            enabled: ${{ inputs.existing }}
+            needs_emulator: false
+          # ppc64le and s390x will become native as soon as we get access to IBM's self-hosted runners
+          - os: ubuntu-24.04
+            arch: ppc64le
+            image: manylinux
+            enabled: ${{ inputs.ibm }}
+            needs_emulator: true  # false
+          - os: ubuntu-24.04
+            arch: s390x
+            image: manylinux
+            enabled: ${{ inputs.ibm }}
+            needs_emulator: true  # false
+          # XXX riscv64 container does not have gn dependency, failing. TODO see if we can procure a pre-build somewhere else.
+          # - os: ubuntu-24.04
+          #   arch: riscv64
+          #   image: manylinux
+          #   enabled: ${{ inputs.emulated }}
+          #   needs_emulator: true
+          # XXX loongarch64 needs patched fork of cibuildwheel, but GH Actions don't support conditional `uses:` expressions. Could use the fork for all, but would prefer not to do that. Could duplicate the cibuildwheel step, or the entire workflow, but neither is nice.
+          # NOTE the loongarch64 container probably won't have gn either, but we can pull it from https://github.com/loong64/gn/releases/latest/download/gn-linux-loong64.tar.gz
+          # - os: ubuntu-24.04
+          #   arch: loongarch64
+          #   image: manylinux
+          #   enabled: ${{ inputs.emulated }}
+          #   needs_emulator: true
+
+    steps:
+
+      - name: Check out the repo
+        if: ${{ matrix.enabled }}
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up QEMU
+        if: ${{ matrix.enabled && matrix.needs_emulator }}
+        uses: docker/setup-qemu-action@v3
+
+      - name: Build wheels
+        if: ${{ matrix.enabled }}
+        uses: pypa/[email protected]  # loong64/[email protected]
+        env:
+          CIBW_ARCHS: ${{ matrix.arch }}
+          # Will be tagged as not python specific by our setup.py. inputs.cibw_py_ver only controls the version used at build time. Could also use `*`, then cibuildwheel would build with the oldest supported version, and walk through the others but skip because a compatible wheel is around already.
+          CIBW_BUILD: "${{ inputs.cibw_py_ver }}-${{ matrix.image }}_${{ matrix.arch }}"
+          # the symlinks are required on aarch64 and loongarch64
+          CIBW_BEFORE_ALL_LINUX: >
+            dnf -y install gn ninja-build freetype-devel glib2-devel lcms2-devel libjpeg-devel libpng-devel libtiff-devel openjpeg2-devel zlib-devel &&
+            ln -s /opt/rh/gcc-toolset-14/root/bin/gcc /opt/rh/gcc-toolset-14/root/bin/${{ matrix.arch }}-linux-gnu-gcc &&
+            ln -s /opt/rh/gcc-toolset-14/root/bin/g++ /opt/rh/gcc-toolset-14/root/bin/${{ matrix.arch }}-linux-gnu-g++ &&
+            ln -s /opt/rh/gcc-toolset-14/root/bin/readelf /opt/rh/gcc-toolset-14/root/bin/${{ matrix.arch }}-linux-gnu-readelf &&
+            ln -s /opt/rh/gcc-toolset-14/root/bin/nm /opt/rh/gcc-toolset-14/root/bin/${{ matrix.arch }}-linux-gnu-nm
+          CIBW_BEFORE_BUILD: python setupsrc/pypdfium2_setup/build_native.py --vendor icu
+          CIBW_ENVIRONMENT: PDFIUM_PLATFORM=sourcebuild
+          # TODO would like to run the test suite too, but not sure how to get the dependencies: CIBW_TEST_REQUIRES assumably takes them from PyPI, which won't work with the exotic architectures. And if we take them from the system with CIBW_BEFORE_TEST, they're not recognized because cibuildwheel sets up a venv or something.
+          CIBW_TEST_COMMAND: python -m pypdfium2 --version
+        with:
+          output-dir: wheelhouse
+
+      - uses: actions/upload-artifact@v4
+        if: ${{ matrix.enabled }}
+        with:
+          path: ./wheelhouse/*.whl
+          name: pypdfium2-wheel-${{ matrix.arch }}

.github/workflows/conda.yaml

@@ -15,7 +15,7 @@ on:
         default: 'latest'
         type: string
       new_only:
-        # only with package == "raw", ignored otherwise (actually the default should be false in that case, but I don't know if GH supports dynamic defaults depending on other inputs)
+        # only with package == "raw", ignored otherwise (actually the default should be false in that case, but don't know if GH supports dynamic defaults depending on other inputs)
         default: true
         type: boolean
       test:

REUSE.toml

@@ -134,6 +134,31 @@ SPDX-FileCopyrightText = [
 ]
 SPDX-License-Identifier = "BSD-3-Clause OR Apache-2.0"
 
+[[annotations]]
+path = "pdfium_patches/cibuildwheel.patch"
+precedence = "aggregate"
+SPDX-FileCopyrightText = [
+    "2025 wojiushixiaobai <[email protected]>",
+    "2025 geisserml <[email protected]>",
+]
+SPDX-License-Identifier = "BSD-3-Clause OR Apache-2.0"
+
+[[annotations]]
+path = [
+    "pdfium_patches/bigendian.patch",
+    "pdfium_patches/bigendian_test.patch",
+]
+precedence = "aggregate"
+SPDX-FileCopyrightText = [
+    "2025 Christian Heimes <[email protected]>",
+    "2025 geisserml <[email protected]>",
+]
+SPDX-License-Identifier = "LicenseRef-Ignore"
+SPDX-FileComment = '''
+Derived from RedHat packaging repository: https://github.com/tiran/libpdfium/blob/d99370b3ac7f0c9cd7222be2dfab2c7b648f2e9e/0001-bigendian.patch
+Repo does not have an own license file, but assumably the patches are being made available under pdfium's license, or leastways an open-source license.
+'''
+
 [[annotations]]
 path = [
     "docs/build/.gitkeep",

pdfium_patches/bigendian.patch

@@ -0,0 +1,40 @@
+diff --git a/core/fxcrt/cfx_seekablestreamproxy.cpp b/core/fxcrt/cfx_seekablestreamproxy.cpp
+index 33cc528..95dc752 100644
+--- a/core/fxcrt/cfx_seekablestreamproxy.cpp
++++ b/core/fxcrt/cfx_seekablestreamproxy.cpp
+@@ -89,11 +89,19 @@ void SwapByteOrder(pdfium::span<uint16_t> str) {
+
+ }  // namespace
+
++#if defined(ARCH_CPU_LITTLE_ENDIAN)
+ #define BOM_UTF8_MASK 0x00FFFFFF
+ #define BOM_UTF8 0x00BFBBEF
+ #define BOM_UTF16_MASK 0x0000FFFF
+ #define BOM_UTF16_BE 0x0000FFFE
+ #define BOM_UTF16_LE 0x0000FEFF
++#else
++#define BOM_UTF8_MASK 0xFFFFFF00
++#define BOM_UTF8 0xEFBBBF00
++#define BOM_UTF16_MASK 0xFFFF0000
++#define BOM_UTF16_BE 0xFEFF0000
++#define BOM_UTF16_LE 0xFFFE0000
++#endif
+
+ CFX_SeekableStreamProxy::CFX_SeekableStreamProxy(
+     const RetainPtr<IFX_SeekableReadStream>& stream)
+@@ -188,9 +196,15 @@ size_t CFX_SeekableStreamProxy::ReadBlock(pdfium::span<wchar_t> buffer) {
+     size_t bytes_read =
+         ReadData(pdfium::as_writable_bytes(buffer).first(bytes_to_read));
+     size_t elements = bytes_read / sizeof(uint16_t);
++#if defined(ARCH_CPU_LITTLE_ENDIAN)
+     if (code_page_ == FX_CodePage::kUTF16BE) {
+       SwapByteOrder(fxcrt::reinterpret_span<uint16_t>(buffer).first(elements));
+     }
++#else
++    if (code_page_ == FX_CodePage::kUTF16LE) {
++      SwapByteOrder(fxcrt::reinterpret_span<uint16_t>(buffer).first(elements));
++    }
++#endif
+     UTF16ToWChar(buffer.first(elements));
+     return elements;
+   }

pdfium_patches/bigendian_test.patch

@@ -0,0 +1,29 @@
+--- a/core/fxcrt/binary_buffer_unittest.cpp
++++ b/core/fxcrt/binary_buffer_unittest.cpp
+@@ -3,6 +3,7 @@
+ // found in the LICENSE file.
+
+ #include "core/fxcrt/binary_buffer.h"
++#include "core/fxcrt/byteorder.h"
+
+ #include <utility>
+ #include <vector>
+@@ -122,7 +123,7 @@
+ // Assumes little endian.
+ TEST(BinaryBuffer, AppendUint16) {
+   BinaryBuffer buffer;
+-  buffer.AppendUint16(0x4321);
++  buffer.AppendUint16(FromLE16(0x4321));
+   EXPECT_EQ(2u, buffer.GetSize());
+   EXPECT_EQ(2u, buffer.GetLength());
+   EXPECT_EQ(0x21u, buffer.GetSpan()[0]);
+@@ -132,7 +133,7 @@
+ // Assumes little endian.
+ TEST(BinaryBuffer, AppendUint32) {
+   BinaryBuffer buffer;
+-  buffer.AppendUint32(0x87654321);
++  buffer.AppendUint32(FromLE32(0x87654321));
+   EXPECT_EQ(4u, buffer.GetSize());
+   EXPECT_EQ(4u, buffer.GetLength());
+   EXPECT_EQ(0x21u, buffer.GetSpan()[0]);
+

pdfium_patches/cibuildwheel.patch

@@ -0,0 +1,29 @@
+diff --git a/core/fxcodec/jpx/cjpx_decoder.cpp b/core/fxcodec/jpx/cjpx_decoder.cpp
+index 4650609..1eadfd5 100644
+--- a/core/fxcodec/jpx/cjpx_decoder.cpp
++++ b/core/fxcodec/jpx/cjpx_decoder.cpp
+@@ -493,9 +493,11 @@ bool CJPX_Decoder::Init(pdfium::span<const uint8_t> src_data,
+   }
+
+   // For https://crbug.com/42270564
++  /* OpenJPEG 2.5 API
+   if (!strict_mode) {
+     CHECK(opj_decoder_set_strict_mode(codec_.get(), false));
+   }
++  */
+
+   opj_image_t* pTempImage = nullptr;
+   if (!opj_read_header(stream_.get(), codec_.get(), &pTempImage)) {
+diff --git a/core/fxge/cfx_face.cpp b/core/fxge/cfx_face.cpp
+index fa02d5b..bfa21d3 100644
+--- a/core/fxge/cfx_face.cpp
++++ b/core/fxge/cfx_face.cpp
+@@ -659,7 +659,7 @@ int CFX_Face::GetCharIndex(uint32_t code) {
+ }
+
+ int CFX_Face::GetNameIndex(const char* name) {
+-  return FT_Get_Name_Index(GetRec(), name);
++  return FT_Get_Name_Index(GetRec(), const_cast<char*>(name));
+ }
+
+ FX_RECT CFX_Face::GetCharBBox(uint32_t code, int glyph_index) {

setupsrc/pypdfium2_setup/base.py

@@ -268,8 +268,7 @@ def to_full(self, v_short):
 
     @cached_property
     def pinned(self):
-        # comments are not permitted in JSON, so the reason for the post_pdfium pin (if set) goes here:
-        # 7309 is the latest tested version. 6996 is too old because we use FPDFFormObj_RemoveObject() which first arrived in 7191.
+        # comments are not permitted in JSON, so the reason for the post_pdfium pin (if set) goes here: (currently no post_pdfium pin)
         record = read_json(AR_RecordFile)
         return record["post_pdfium"] or record["pdfium"]