From e871e928b3a95a88244c6b09c13e027137d4d9f5 Mon Sep 17 00:00:00 2001
From: Mauro Druwel <46003176+MauroDruwel@users.noreply.github.com>
Date: Sat, 9 Aug 2025 19:46:33 +0200
Subject: [PATCH] Update install.sh

---
 install.sh | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/install.sh b/install.sh
index 43ccd2d..8a8c428 100644
--- a/install.sh
+++ b/install.sh
@@ -43,8 +43,23 @@ if [ "${systemName}" = "Raspberry Pi" ] || [ "${systemName}" = "Orange Pi" ] ||
     installAptPackages "libatlas-base-dev libopenblas-dev libblas-dev"
     installAptPackages "liblapack-dev patchelf gfortran"
 
+    # Fix for lxml build failures on aarch64 when installing PaddleOCR dependencies.
+    # lxml>=5/6 may be built from source if no wheel is available on the mirror.
+    # Ensure system headers exist so a source build can succeed, and preinstall a
+    # compatible binary lxml for Python 3.8 to avoid building from source.
+    installAptPackages "libxml2-dev libxslt1-dev zlib1g-dev pkg-config"
+
     if [ "$oneStepPIP" = false ]; then
         installPythonPackagesByName "Cython six requests wheel pyyaml"
+        # Prefer a prebuilt wheel for Python 3.8 to avoid source builds on aarch64 mirrors.
+        if [ "${pythonVersion}" = "3.8" ]; then
+            installPythonPackagesByName "lxml==4.9.4" "lxml XML/HTML parser (preinstalled to avoid source build)"
+        fi
+    else
+        # When doing one-step pip, still try to mitigate lxml builds for Py3.8
+        if [ "${pythonVersion}" = "3.8" ]; then
+            installPythonPackagesByName "lxml==4.9.4" "lxml XML/HTML parser (preinstalled to avoid source build)"
+        fi
     fi
 
     "$venvPythonCmdPath" -m pip show paddlepaddle >/dev/null 2>/dev/null
@@ -69,6 +84,17 @@ if [ "${systemName}" = "Raspberry Pi" ] || [ "${systemName}" = "Orange Pi" ] ||
 
 elif [ "${systemName}" = "Jetson" ]; then
 
+    # Ensure system headers for lxml in case a source build is needed by PaddleOCR deps
+    write "Ensuring system headers for lxml are present..." $color_info
+    sudo apt-get update -y >/dev/null 2>/dev/null
+    installAptPackages "libxml2-dev libxslt1-dev zlib1g-dev pkg-config"
+    writeLine "Done" $color_success
+
+    # For Python 3.8 on Jetson, preinstall a binary lxml to avoid building from source on aarch64
+    if [ "${pythonVersion}" = "3.8" ]; then
+        installPythonPackagesByName "lxml==4.9.4" "lxml XML/HTML parser (preinstalled to avoid source build)"
+    fi
+
     case "$pythonVersion" in
         "3.6") 
             wheel_file="paddlepaddle_gpu-2.4.1-cp36-cp36m-linux_aarch64.whl"
@@ -159,6 +185,11 @@ if [ "$os" = "macos" ]; then
     else
         installPythonPackagesByName "opencv-python"           "OpenCV, the Computer Vision library for Python"
     fi
+
+    # Preinstall lxml to avoid source builds on macOS if PaddleOCR pulls it in
+    if [ "${pythonVersion}" = "3.8" ]; then
+        installPythonPackagesByName "lxml==4.9.4" "lxml XML/HTML parser (preinstalled to avoid source build)"
+    fi
 fi
 
 if [ "${moduleInstallErrors}" = "" ]; then