wasm: Parse wasm binaries to get dependencies

Currently we are searching binaries for string ending with ".so" in
order to find shared library dependencies. This is not the cleanest
solution.
Parse wasm binaries instead, and extract the information in structured
way.

Fixes: QTBUG-121833
Change-Id: Id0efb07af4efe7641ed4fd66f76c395722478f95
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>

2 files changed, 82 insertions, 11 deletions

diff --git a/util/wasm/preload/preload_qml_imports.py b/util/wasm/preload/preload_qml_imports.py
index 2f2fc7958f..42fea6e087 100755
--- a/util/wasm/preload/preload_qml_imports.py
+++ b/util/wasm/preload/preload_qml_imports.py
@@ -8,6 +8,8 @@ import subprocess
 import json
 import re
 
+from wasm_binary_tools import WasmBinary
+
 # Paths to shared libraries and qml imports on the Qt installation on the web server.
 # "$QTDIR" is replaced by qtloader.js at load time (defaults to "qt"), and makes
 # possible to relocate the application build relative to the Qt build on the web server.
@@ -29,17 +31,8 @@ def preload_file(source, destination):
 
 
 def find_dependencies(filepath):
-    # Very basic dependency finder which scans for ".so" strings in the file
-    try:
-        with open(filepath, "rb") as file:
-            content = file.read()
-            return [
-                m.group(0).decode("utf-8")
-                for m in re.finditer(rb"[\w\-.]+\.so", content)
-            ]
-    except IOError as e:
-        eprint(f"Error: {e}")
-        return []
+    binary = WasmBinary(filepath)
+    return binary.get_dependencies()
 
 
 def extract_preload_files_from_imports(imports):
diff --git a/util/wasm/preload/wasm_binary_tools.py b/util/wasm/preload/wasm_binary_tools.py
new file mode 100644
index 0000000000..9a2150b964
--- /dev/null
+++ b/util/wasm/preload/wasm_binary_tools.py
@@ -0,0 +1,78 @@
+
+#!/usr/bin/env python3
+# Copyright (C) 2024 The Qt Company Ltd.
+# SPDX-License-Identifier: LicenseRef-Qt-Commercial OR GPL-3.0-only WITH Qt-GPL-exception-1.0
+
+import sys
+import struct
+
+
+class WasmBinary:
+    """For reference of binary format see Emscripten source code, especially library_dylink.js."""
+
+    def __init__(self, filepath):
+        self._offset = 0
+        self._end = 0
+        self._dependencies = []
+        with open(filepath, 'rb') as file:
+            self._binary = file.read()
+            self._check_preamble()
+            self._parse_subsections()
+
+    def get_dependencies(self):
+        return self._dependencies
+
+    def _get_leb(self):
+        ret = 0
+        mul = 1
+        while True:
+            byte = self._binary[self._offset]
+            self._offset += 1
+            ret += (byte & 0x7f) * mul
+            mul *= 0x80
+            if not (byte & 0x80):
+                break
+        return ret
+
+    def _get_string(self):
+        length = self._get_leb()
+        self._offset += length
+        return self._binary[self._offset - length:self._offset].decode('utf-8')
+
+    def _check_preamble(self):
+        preamble = memoryview(self._binary)[:24]
+        int32View = struct.unpack('<6I', preamble)
+        assert int32View[0] == 0x6d736100, "magic number not found"
+        assert self._binary[8] == 0, "dynlink section needs to be first"
+        self._offset = 9
+        section_size = self._get_leb()
+        self._end = self._offset + section_size
+        name = self._get_string()
+        assert name == "dylink.0", "section dylink.0 not found"
+
+    def _parse_subsections(self):
+        WASM_DYLINK_NEEDED = 0x2
+
+        while self._offset < self._end:
+            subsection_type = self._binary[self._offset]
+            self._offset += 1
+            subsection_size = self._get_leb()
+
+            if subsection_type == WASM_DYLINK_NEEDED:
+                needed_dynlibs_count = self._get_leb()
+                for _ in range(needed_dynlibs_count):
+                    self._dependencies.append(self._get_string())
+            else:
+                self._offset += subsection_size  # we don't care about other sections for now
+
+
+if __name__ == "__main__":
+    if len(sys.argv) != 2:
+        print("Usage: python wasm_binary_tools.py <shared_object>")
+        sys.exit(1)
+
+    file_path = sys.argv[1]
+    binary = WasmBinary(file_path)
+    dependencies = binary.get_dependencies()
+    for d in dependencies:
+        print(d)