From 1e283fd3e071d5085d175429929662df3ed70fc5 Mon Sep 17 00:00:00 2001
From: falsycat <me@falsy.cat>
Date: Sun, 27 Aug 2023 10:10:50 +0900
Subject: [PATCH] create new project

---
 .gitignore     |   3 ++
 README.md      | 101 ++++++++++++++++++++++++++++++++++++
 castre.py      | 136 +++++++++++++++++++++++++++++++++++++++++++++++++
 pyproject.toml |   6 +++
 4 files changed, 246 insertions(+)
 create mode 100644 .gitignore
 create mode 100644 README.md
 create mode 100644 castre.py
 create mode 100644 pyproject.toml

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..cb37f5c
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,3 @@
+/__pycache__/
+/castre.egg-info/
+/build/
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..8763549
--- /dev/null
+++ b/README.md
@@ -0,0 +1,101 @@
+castre
+====
+
+This is a python library for C++ AST-based REfactoring
+
+- easy
+- simple
+- minimal
+- changes only where to refactor
+
+## How to Use
+
+1. install this package
+1. write a script to refactor like an example below
+1. execute the script
+1. execute your formatter
+1. verify and approve the changes by `git add -p`
+1. save the approved changes by `git commit -m ...`
+1. discard the unapproved changes by `git restore .`
+
+## Example
+
+refactoring script:
+```python
+import castre
+import re
+
+def walker(item):
+  # The actual AST json data is stored at `item.raw`
+  # You can see it by `clang++ -Xclang -ast-dump=json -c [filename and options...]`
+  # And you can also see an AST structure in human-readable form by -ast-dump option without any value
+  # Please note that outputs of both commands are huge
+  if item.raw["kind"] == "DeclStmt" and item.raw["inner"][0]["kind"] == "VarDecl":
+    item.refactor(fix)
+  else:
+    for child in item:
+      walker(child)
+
+def fix(text):
+  if text.startswith("auto"):
+    text = "\n// meta comment to ignore violation in the next line\n" + text
+  else:
+    text = re.sub(r"^([^=]*)=(.*)$", r"\1{\2}", text, flags=re.S)
+  return text
+
+# parse cpp codes and queue refactoring tasks
+fixer = castre.walk(
+    ["a.cc", "-I.", "-std=c++20"],
+    walker,
+    path_filter=lambda x: x is not None and x.startswith("/Users/falsycat"))
+
+# you can reuse the fixer
+castre.walk(
+    ["b.cc", "-I.", "-std=c++20"],
+    walker,
+    fixer=fixer,
+    path_filter=lambda x: x is not None and x.startswith("/Users/falsycat"))
+
+# execute the tasks and apply changes to the actual files
+fixer.fix()
+```
+
+before:
+```cpp
+#include <iostream>
+
+int main() {
+  struct A { void f() { int a = 123; auto b = 1+2+3+4; } };
+  int x = 123;
+std::cout << "helloworld"
+<< std::endl;
+  auto y = 123;
+    std::cout << "goodbye" << std::endl;
+  int
+  z
+  =
+  123;
+}
+```
+
+after:
+```cpp
+#include <iostream>
+
+int main() {
+  struct A { void f() { int a { 123}; 
+// meta comment to ignore violation in the next line
+auto b = 1+2+3+4; } };
+  int x { 123};
+std::cout << "helloworld"
+<< std::endl;
+  
+// meta comment to ignore violation in the next line
+auto y = 123;
+    std::cout << "goodbye" << std::endl;
+  int
+  z
+  {
+  123};
+}
+```
diff --git a/castre.py b/castre.py
new file mode 100644
index 0000000..d3b7726
--- /dev/null
+++ b/castre.py
@@ -0,0 +1,136 @@
+import bisect
+import json
+import os
+import subprocess
+
+def walk(args, walker, path_filter=None, fixer=None):
+  proc = subprocess.run(["clang++", "-Xclang", "-ast-dump=json", "-c", *args], capture_output=True)
+  if proc.returncode != 0:
+    print(proc.stderr.decode("utf-8"))
+    raise Exception("analysis failure")
+
+  tree =  json.loads(proc.stdout)
+  root = Item(tree)
+
+  fixer = fixer if fixer is not None else Fixer()
+  file  = None
+  path  = None
+  for item in tree["inner"]:
+    if "file" in item["loc"]:
+      path = os.path.abspath(item["loc"]["file"])
+      file = None
+
+    if path_filter is not None and not path_filter(path):
+      continue
+
+    if file is None:
+      file = fixer.makeFile(path)
+
+    walker(Item(item, parent=root, file=file))
+
+  return fixer
+
+class Item:
+  def __init__(self, j, parent=None, file=None):
+    self.parent = parent
+    self.raw    = j
+    self.file   = None
+
+    if parent is not None:
+      self.file = parent.file
+    if file is not None:
+      self.file = file
+
+  def __iter__(self):
+    return ItemItr(self)
+
+  def refactorable(self):
+    return self.file is not None and  "file" not in self.raw["range"]["end"]
+
+  def range(self):
+    if not self.refactorable():
+      raise Exception("item is not refactorable")
+
+    ra    = self.raw["range"]
+    begin = ra["begin"]["offset"]
+    end   = ra["end"]["offset"]
+    return (begin, end)
+
+  def pos(self):
+    begin, end = self.range()
+    return (begin, end-begin)
+
+  def refactor(self, text):
+    self.file.replace(*self.pos(), text)
+
+  def insertBefore(self, text):
+    self.file.replace(range()[0], 0, text)
+
+  def insertAfter(self, text):
+    self.file.replace(range()[1], 0, text)
+
+class ItemItr:
+  def __init__(self, item):
+    self.index = 0
+    self.item  = item
+    self.inner = self.item.raw["inner"] if "inner" in self.item.raw else None
+
+  def __next__(self):
+    if self.inner == None or len(self.inner) <= self.index:
+      raise StopIteration()
+    ret = Item(self.inner[self.index], parent=self.item)
+    self.index += 1
+    return ret
+
+class Fixer:
+  def __init__(self):
+    self.files = {}
+
+  def makeFile(self, path):
+    if path in self.files:
+      return self.files[path]
+    ret = File(path);
+    self.files[path] = ret
+    return ret
+
+  def fix(self):
+    for f in self.files.values():
+      f.fix()
+
+class File:
+  def __init__(self, path):
+    self.path  = path
+    self.tasks = []
+
+  def replace(self, offset, n, text):
+    idx = bisect.bisect(self.tasks, x=offset, key=lambda x: x[0])
+
+    if idx > 0:
+      prev = self.tasks[idx-1]
+      if prev[0]+prev[1] > offset:
+        raise Exception("change conflict")
+
+    if idx < len(self.tasks):
+      next = self.tasks[idx]
+      if next[0] < offset+n:
+        raise Exception ("change conflict")
+
+    self.tasks.insert(idx, (offset, n, text))
+
+  def dryFix(self):
+    with open(self.path, "r") as f:
+      src = f.read()
+
+    for task in reversed(self.tasks):
+      begin = task[0]
+      end   = begin + task[1]
+      text  = task[2]
+      if not isinstance(text, str):
+        text = text(src[begin:end])
+      src = src[0:begin] + text + src[end:]
+    return src
+
+  def fix(self):
+    src = self.dryFix()
+    with open(self.path, "w") as f:
+      f.write(src)
diff --git a/pyproject.toml b/pyproject.toml
new file mode 100644
index 0000000..18553e0
--- /dev/null
+++ b/pyproject.toml
@@ -0,0 +1,6 @@
+[project]
+name = "castre"
+version = "0.9"
+
+[project.scripts]
+castre = "castre:castre"