Issue 108: Remove logic to find LLVM in linux OSes. Only appears in Darwin. And also add some test cases.

Signed-off-by: Rahul Krishna <[email protected]>

Author: rahlk

Makefile

@@ -31,7 +31,7 @@ lint: ## Run the linter
 .PHONY: test
 test: ## Run the unit tests
 	$(info Running tests...)
-	pytest --pspec --cov=cldk --cov-fail-under=50 --disable-warnings
+	pytest --pspec --cov=cldk --cov-fail-under=75 --disable-warnings
 
 ##@ Build
 

cldk/analysis/c/c_analysis.py

@@ -18,11 +18,11 @@
 Analysis model for C projects
 """
 
+import os
 from pathlib import Path
 from typing import Dict, List, Optional
 import networkx as nx
 
-
 from cldk.analysis.c.clang import ClangAnalyzer
 from cldk.models.c import CApplication, CFunction, CTranslationUnit, CMacro, CTypedef, CStruct, CEnum, CVariable
 
@@ -36,7 +36,7 @@ def __init__(self, project_dir: Path) -> None:
         self.c_application = self._init_application(project_dir)
 
     def _init_application(self, project_dir: Path) -> CApplication:
-        """Initializes the C application object.
+        """Should initialize the C application object.
 
         Args:
             project_dir (Path): Path to the project directory.

cldk/analysis/c/clang/clang_analyzer.py

@@ -1,4 +1,5 @@
 import os
+from pdb import set_trace
 import platform
 from clang.cindex import Config
 from pathlib import Path
@@ -19,16 +20,27 @@ class ClangAnalyzer:
     """Analyzes C code using Clang's Python bindings."""
 
     def __init__(self, compilation_database_path: Optional[Path] = None):
+        # # Let's turn off Address sanitization for parsing code
+        # # Initialize libclang at module level
+        # try:
+        if platform.system() == "Darwin":
+            possible_paths = [
+                "/opt/homebrew/opt/llvm/lib/libclang.dylib",  # Apple Silicon
+                "/usr/local/opt/llvm/lib/libclang.dylib",  # Intel Mac
+                "/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/lib/libclang.dylib",
+            ]
 
-        # Initialize libclang at module level
-        try:
-            Config.set_library_file(self.__find_libclang())
-        except Exception as e:
-            logger.error(f"Failed to initialize libclang: {e}")
-            raise Exception("Failed to initialize libclang")
+            # We could not find libclang. Raise an error and provide instructions.
+            if len(possible_paths) == 0:
+                raise RuntimeError("Install LLVM 18 using: brew install llvm@18")
+
+            # Check each possible path and return the first one that exists
+            for path in possible_paths:
+                if os.path.exists(path):
+                    logger.info(f"Found libclang at: {path}")
+                    # Configure Clang before creating the Index
+                    Config.set_library_file(path)
 
-        logger.info("Successfully initialized libclang")
-        # Configure Clang before creating the Index
         self.index = Index.create()
         self.compilation_database = None
         # TODO: Implement compilation database for C/C++ projects so that we can get compile arguments for each file
@@ -59,8 +71,8 @@ def __find_libclang(self) -> str:
             from pathlib import Path
 
             lib_paths = [Path("/usr/lib"), Path("/usr/lib64")]
-            possible_paths = [str(p) for base in lib_paths if base.exists() for p in base.rglob("libclang*.so*")]
-
+            possible_paths = [str(p) for base in lib_paths if base.exists() for p in base.rglob("libclang*.so.17*")]
+            print(possible_paths)
             install_instructions = "Install libclang development package using your system's package manager"
         else:
             raise RuntimeError(f"Unsupported operating system: {system}")
@@ -98,7 +110,7 @@ def analyze_file(self, file_path: Path) -> CTranslationUnit:
         return translation_unit
 
     def _process_translation_unit(self, cursor, translation_unit: CTranslationUnit):
-        """Processes all declarations in a translation unit."""
+        """Should process all declarations in a translation unit."""
 
         for child in cursor.get_children():
             if child.location.file and str(child.location.file) != translation_unit.file_path:

cldk/analysis/java/codeanalyzer/codeanalyzer.py

@@ -153,7 +153,7 @@ def _get_codeanalyzer_exec(self) -> List[str]:
 
     @staticmethod
     def _init_japplication(data: str) -> JApplication:
-        """return JApplication giving the stringified JSON as input.
+        """Should return JApplication giving the stringified JSON as input.
         Returns
         -------
         JApplication
@@ -165,7 +165,7 @@ def _init_japplication(data: str) -> JApplication:
         return JApplication(**json.loads(data))
 
     def _init_codeanalyzer(self, analysis_level=1) -> JApplication:
-        """Initializes the Codeanalyzer.
+        """Should initialize the Codeanalyzer.
 
         Args:
             analysis_level (int): The level of analysis to be performed (1 for symbol table, 2 for call graph).
@@ -657,7 +657,7 @@ def get_class_call_graph_using_symbol_table(self, qualified_class_name: str, met
         return graph_edges
 
     def __call_graph_using_symbol_table(self, qualified_class_name: str, method_signature: str, is_target_method: bool = False) -> nx.DiGraph:
-        """Generate call graph using symbol table
+        """Should generate call graph using symbol table
         Args:
             qualified_class_name: qualified class name
             method_signature: method signature

cldk/analysis/java/codeql/codeql.py

@@ -74,7 +74,7 @@ def __init__(self, project_dir: Union[str, Path], codeql_db: Union[str, Path, No
 
     @staticmethod
     def _init_codeql_db(project_dir: Union[str, Path], codeql_db: Union[str, Path, None]) -> Path:
-        """Initializes the CodeQL database.
+        """Should initialize the CodeQL database.
 
         Parameters
         ----------

cldk/analysis/java/treesitter/java_sitter.py

@@ -229,7 +229,7 @@ def safe_ascend(self, node: Node, ascend_count: int) -> Node:
             return self.safe_ascend(node.parent, ascend_count - 1)
 
     def get_call_targets(self, method_body: str, declared_methods: dict) -> Set[str]:
-        """Generate a list of call targets from the method body.
+        """Should generate a list of call targets from the method body.
 
         Uses simple name resolution for finding the call targets. Nothing sophisticated here. Just a simple search
         over the AST.

cldk/models/treesitter/models.py

@@ -66,11 +66,11 @@ def __getitem__(self, index: int) -> Capture:
         return self.captures[index]
 
     def __iter__(self):
-        """return an iterator over the captures."""
+        """Should return an iterator over the captures."""
         return iter(self.captures)
 
     def __len__(self) -> int:
-        """return the number of captures."""
+        """Should return the number of captures."""
         return len(self.captures)
 
     def __add__(self, other: "Captures") -> "Captures":