[Lldb-commits] [lldb] [lldb][RPC] Upstream Python scripts (PR #138028)

[Chelsea Cassanova via lldb-commits]

================
@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+# Usage: convert-lldb-header-to-rpc-header.py <path/to/input-header.h> <path/to/output-header.h>
+# This scripts takes common LLDB headers (such as lldb-defines.h) and replaces references to LLDB
+# with those for RPC. This happens for:
+# - namespace definitions
+# - namespace usage
+# - version string macros
+# - ifdef/ifndef lines
+
+import argparse
+import os
+import re
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("input")
+    parser.add_argument("output")
+    args = parser.parse_args()
+    input_path = str(args.input)
+    output_path = str(args.output)
+    with open(input_path, "r") as input_file:
+        lines = input_file.readlines()
+
+    with open(output_path, "w") as output_file:
+        for line in lines:
+            # NOTE: We do not use lldb-forward.h or lldb-versioning.h in RPC, so remove
+            # all includes that are found for these files.
+            if re.match(
+                r'#include "lldb/lldb-forward|#include "lldb/lldb-versioning', line
+            ):
+                continue
+            # For lldb-rpc-defines.h, replace the ifndef LLDB_LLDB_ portion with LLDB_RPC_ as we're not
+            # using LLDB private definitions in RPC.
+            elif re.match(r".+LLDB_LLDB_", line):
+                output_file.write(re.sub(r"LLDB_LLDB_", r"LLDB_RPC_", line))
+            # Similarly to lldb-rpc-defines.h, replace the ifndef for LLDB_API in SBDefines.h to LLDB_RPC_API_ for the same reason.
+            elif re.match(r".+LLDB_API_", line):
+                output_file.write(re.sub(r"LLDB_API_", r"LLDB_RPC_API_", line))
----------------
chelcassanova wrote:

I looked at running only the `sub` and I think I'll actually keep it as is here. The reason for that is because the general pattern here if "if we need to change the line then change it, otherwise just write whatever was in the original input file". Running only the `sub` without if clauses makes it harder to know when we need to just use the original input file line. I'd have preferred for this to be a `switch` statement (or `case` in Python's case) but IIUC the current version of Python that we support hasn't caught up to using that statement yet.

https://github.com/llvm/llvm-project/pull/138028