From 06615bdb1969d9f13d65cbbc755e1303dbe35d6a Mon Sep 17 00:00:00 2001
From: Jose Rodriguez <boriel@gmail.com>
Date: Sat, 29 Mar 2025 12:43:23 +0100
Subject: [PATCH] feat: adds a tool to consolidate (hard-link) duplicated files

Many files in the libs/ directory are duplicated because
most archs share a lot in common. Mantaining modification along
all possible archs is tedious, so this tool hardlinks all
identical files son modifying one is modifying all of them.

Git is agnostic to this.
---
 tools/consolidate-libs.py | 61 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 61 insertions(+)
 create mode 100755 tools/consolidate-libs.py

diff --git a/tools/consolidate-libs.py b/tools/consolidate-libs.py
new file mode 100755
index 000000000..4ca91ca67
--- /dev/null
+++ b/tools/consolidate-libs.py
@@ -0,0 +1,61 @@
+#!/usr/bin/env python
+
+__doc__ = """Scans src/lib/<arch>/** and does hardlinks to files
+with the same name and content"""
+
+import glob
+import os
+import re
+from collections import defaultdict
+from pathlib import Path
+from typing import NamedTuple
+
+ROOT_DIR = Path(__file__).parent.parent.absolute() / "src" / "lib" / "arch"
+ARCHS = "zx48k", "zxnext"
+
+
+class FileInfo(NamedTuple):
+    path: str
+    hash: int
+
+
+def get_file_list(root: Path) -> list[str]:
+    filelist = glob.glob(str(root / "**" / "*"), recursive=True)
+    return [f for f in filelist if os.path.isfile(f)]
+
+
+def scan_arch(root: Path) -> dict[FileInfo, list[str]]:
+    result = defaultdict(list)
+    re_arch = re.compile(r"^.*?/src/lib/arch/[^/]+/(.*)$")
+
+    files = get_file_list(root)
+    for file in files:
+        match = re_arch.match(file)
+        if not match:
+            continue
+
+        path = match.group(1)
+        result[FileInfo(path=path, hash=hash(open(file, "rb").read()))].append(file)
+
+    return result
+
+
+def fold_files(scan: dict[FileInfo, list[str]]) -> None:
+    for path, files in scan.items():
+        if len(files) == 1:
+            continue
+
+        main_file = files[0]
+        for file in files[1:]:
+            print(f"Linking {main_file} to {file}")
+            os.unlink(file)
+            os.link(main_file, file)
+
+
+def main():
+    scan = scan_arch(ROOT_DIR)
+    fold_files(scan)
+
+
+if __name__ == "__main__":
+    main()