Create temporary tables to improve performance

Makefile

@@ -23,5 +23,7 @@ fmt:              ## Format code using black & isort.
 lint:             ## Run pep8, black, mypy linters.
 	flake8 sqlelf/
 	black --check sqlelf/
-	pyright
+# TODO(fzakaria): without pythonpath it picks up the wrong python
+# and then does not find the venv for the imports
+	pyright --pythonpath $(shell which python)
 	nixpkgs-fmt --check .

flake.nix

@@ -27,6 +27,6 @@
         };
 
         devShell = pkgs.sqlelf-env.env.overrideAttrs
-          (oldAttrs: { buildInputs = with pkgs; [ poetry pyright nixpkgs-fmt ]; });
+          (oldAttrs: { buildInputs = with pkgs; [ poetry nixpkgs-fmt ]; });
       });
 }

pyproject.toml

@@ -21,6 +21,7 @@ setuptools = "*"
 black = "^23.7.0"
 isort = "^5.12.0"
 flake8 = "^6.1.0"
+pyright = "^1.1.325"
 
 [build-system]
 requires = ["poetry-core"]

sqlelf/elf/instruction.py

@@ -63,4 +63,10 @@ def register(connection: apsw.Connection, binaries: list[lief.Binary]):
     generator.columns, generator.column_access = apsw.ext.get_column_names(
         next(generator())
     )
-    apsw.ext.make_virtual_module(connection, "elf_instructions", generator)
+    apsw.ext.make_virtual_module(connection, "raw_elf_instructions", generator)
+    connection.execute(
+        """
+        CREATE TEMP TABLE elf_instructions
+        AS SELECT * FROM raw_elf_instructions;
+        """
+    )

sqlelf/elf/section.py

@@ -27,6 +27,12 @@ def generator() -> Iterator[dict[str, Any]]:
     return generator
 
 
+def section_name(name: str | None) -> str | None:
+    if name == "":
+        return "undefined"
+    return name
+
+
 def register(connection: apsw.Connection, binaries: list[lief.Binary]):
     generator = elf_sections(binaries)
     # setup columns and access by providing an example of the first entry returned

sqlelf/elf/symbol.py

@@ -6,6 +6,7 @@
 import apsw
 import apsw.ext
 import lief
+from ..elf.section import section_name as elf_section_name
 
 
 def elf_symbols(binaries: list[lief.Binary]):
@@ -15,6 +16,16 @@ def generator() -> Iterator[dict[str, Any]]:
             # as they can be costly
             binary_name = binary.name
             for symbol in binary.symbols:
+                # The section index can be special numbers like 65521 or 65522
+                # that refer to special sections so they can't be indexed
+                section_name: str | None = next(
+                    (
+                        section.name
+                        for shndx, section in enumerate(binary.sections)
+                        if shndx == symbol.shndx
+                    ),
+                    None,
+                )
                 yield {
                     "path": binary_name,
                     "name": symbol.name,
@@ -30,7 +41,7 @@ def generator() -> Iterator[dict[str, Any]]:
                     # https://www.m4b.io/elf/export/binary/analysis/2015/05/25/what-is-an-elf-export.html
                     "imported": symbol.imported,
                     "exported": symbol.exported,
-                    "section": binary.sections[symbol.shndx].name,
+                    "section": elf_section_name(section_name),
                     "size": symbol.size,
                 }
 
@@ -43,4 +54,10 @@ def register(connection: apsw.Connection, binaries: list[lief.Binary]):
     generator.columns, generator.column_access = apsw.ext.get_column_names(
         next(generator())
     )
-    apsw.ext.make_virtual_module(connection, "elf_symbols", generator)
+    apsw.ext.make_virtual_module(connection, "raw_elf_symbols", generator)
+    connection.execute(
+        """
+        CREATE TEMP TABLE elf_symbols
+        AS SELECT * FROM raw_elf_symbols;
+        """
+    )