Merge branch 'develop' into staging

.github/workflows/ci.yml

@@ -21,7 +21,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: [ '3.7', '3.8', '3.9', '3.10' ]
+        python-version: ['3.8', '3.9', '3.10' ]
 
     name: Base (${{ matrix.python-version }})
 
@@ -58,7 +58,7 @@ jobs:
 
     strategy:
       matrix:
-        python-version: [ '3.7', '3.10' ]
+        python-version: ['3.10' ]
 
     name: Production (${{ matrix.python-version }})
 
@@ -96,7 +96,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: [ '3.7', '3.10' ]
+        python-version: ['3.10' ]
 
     name: Tests (${{ matrix.python-version }})
 
@@ -120,11 +120,6 @@ jobs:
         with:
           node-version: 18
 
-      - uses: actions/setup-node@v3
-        if: ${{ matrix.python-version == '3.7' }}
-        with:
-          node-version: 14
-
       - run: |
           wget https://github.com/wkhtmltopdf/packaging/releases/download/0.12.6-1/wkhtmltox_0.12.6-1.focal_amd64.deb;
           sudo apt install ./wkhtmltox_0.12.6-1.focal_amd64.deb;

README.md

@@ -10,7 +10,7 @@ Bench is a command-line utility that helps you to install, update, and manage mu
 
 <div align="center">
 	<a target="_blank" href="https://www.python.org/downloads/" title="Python version">
-		<img src="https://img.shields.io/badge/python-%3E=_3.7-green.svg">
+		<img src="https://img.shields.io/badge/python-%3E=_3.8-green.svg">
 	</a>
 	<a target="_blank" href="https://app.travis-ci.com/github/frappe/bench" title="CI Status">
 		<img src="https://app.travis-ci.com/frappe/bench.svg?branch=develop">

bench/app.py

@@ -6,6 +6,7 @@ import re
 import shutil
 import subprocess
 import sys
+import uuid
 import tarfile
 import typing
 from collections import OrderedDict
@@ -34,6 +35,7 @@ from bench.utils import (
 	is_valid_frappe_branch,
 	log,
 	run_frappe_cmd,
+	get_file_md5,
 )
 from bench.utils.bench import build_assets, install_python_dev_dependencies
 from bench.utils.render import step
@@ -338,45 +340,51 @@ class App(AppMeta):
 	def get_app_path(self) -> Path:
 		return Path(self.bench.name) / "apps" / self.app_name
 
-	def get_app_cache_path(self, is_compressed=False) -> Path:
+	def get_app_cache_temp_path(self, is_compressed=False) -> Path:
+		cache_path = get_bench_cache_path("apps")
+		ext = "tgz" if is_compressed else "tar"
+		tarfile_name = f"{self.app_name}.{uuid.uuid4().hex}.{ext}"
+		return cache_path / tarfile_name
+
+	def get_app_cache_hashed_path(self, temp_path: Path) -> Path:
 		assert self.cache_key is not None
 
-		cache_path = get_bench_cache_path("apps")
+		ext = temp_path.suffix[1:]
-		tarfile_name = get_cache_filename(
+		md5 = get_file_md5(temp_path)
-			self.app_name,
+		tarfile_name = f"{self.app_name}.{self.cache_key}.md5-{md5}.{ext}"
-			self.cache_key,
+
-			is_compressed,
+		return temp_path.with_name(tarfile_name)
-		)
-		return cache_path / tarfile_name
 
 	def get_cached(self) -> bool:
 		if not self.cache_key:
 			return False
 
-		cache_path = self.get_app_cache_path(False)
+		if not (cache_path := validate_cache_and_get_path(self.app_name, self.cache_key)):
-		mode = "r"
-
-		# Check if cache exists without gzip
-		if not cache_path.is_file():
-			cache_path = self.get_app_cache_path(True)
-			mode = "r:gz"
-
-		# Check if cache exists with gzip
-		if not cache_path.is_file():
 			return False
 
 		app_path = self.get_app_path()
 		if app_path.is_dir():
 			shutil.rmtree(app_path)
 
-		click.secho(f"Getting {self.app_name} from cache", fg="yellow")
+		click.secho(
+			f"Bench app-cache: extracting {self.app_name} from {cache_path.as_posix()}",
+		)
+
+		mode = "r:gz" if cache_path.suffix.endswith(".tgz") else "r"
 		with tarfile.open(cache_path, mode) as tar:
 			extraction_filter = get_app_cache_extract_filter(count_threshold=150_000)
 			try:
 				tar.extractall(app_path.parent, filter=extraction_filter)
+				click.secho(
+					f"Bench app-cache: extraction succeeded for {self.app_name}",
+					fg="green",
+				)
 			except Exception:
-				message = f"Cache extraction failed for {self.app_name}, skipping cache"
+				message = f"Bench app-cache: extraction failed for {self.app_name}"
-				click.secho(message, fg="yellow")
+				click.secho(
+					message,
+					fg="yellow",
+				)
 				logger.exception(message)
 				shutil.rmtree(app_path)
 				return False
@@ -392,10 +400,10 @@ class App(AppMeta):
 			return False
 
 		cwd = os.getcwd()
-		cache_path = self.get_app_cache_path(compress_artifacts)
+		cache_path = self.get_app_cache_temp_path(compress_artifacts)
 		mode = "w:gz" if compress_artifacts else "w"
 
-		message = f"Caching {self.app_name} app directory"
+		message = f"Bench app-cache: caching {self.app_name}"
 		if compress_artifacts:
 			message += " (compressed)"
 		click.secho(message)
@@ -407,9 +415,19 @@ class App(AppMeta):
 		try:
 			with tarfile.open(cache_path, mode) as tar:
 				tar.add(app_path.name)
+
+			hashed_path = self.get_app_cache_hashed_path(cache_path)
+			unlink_no_throw(hashed_path)
+
+			cache_path.rename(hashed_path)
+			click.secho(
+				f"Bench app-cache: caching succeeded for {self.app_name} as {hashed_path.as_posix()}",
+				fg="green",
+			)
+
 			success = True
-		except Exception:
+		except Exception as exc:
-			log(f"Failed to cache {app_path}", level=3)
+			log(f"Bench app-cache: caching failed for {self.app_name} {exc}", level=3)
 			success = False
 		finally:
 			os.chdir(cwd)
@@ -437,28 +455,11 @@ def can_get_cached(app_name: str, cache_key: str) -> bool:
 	checking local remote and fetching can be skipped while keeping
 	get-app command params the same.
 	"""
-	cache_path = get_bench_cache_path("apps")
-	tarfile_path = cache_path / get_cache_filename(
-		app_name,
-		cache_key,
-		True,
-	)
 
-	if tarfile_path.is_file():
+	if cache_path := get_app_cache_path(app_name, cache_key):
-		return True
+		return cache_path.exists()
 
-	tarfile_path = cache_path / get_cache_filename(
+	return False
-		app_name,
-		cache_key,
-		False,
-	)
-
-	return tarfile_path.is_file()
-
-
-def get_cache_filename(app_name: str, cache_key: str, is_compressed=False):
-	ext = "tgz" if is_compressed else "tar"
-	return f"{app_name}-{cache_key[:10]}.{ext}"
 
 
 def can_frappe_use_cached(app: App) -> bool:
@@ -482,7 +483,10 @@ def can_frappe_use_cached(app: App) -> bool:
 		"""
 		return sv.Version("15.12.0") not in sv.SimpleSpec(min_frappe)
 	except ValueError:
-		click.secho(f"Invalid value found for frappe version '{min_frappe}'", fg="yellow")
+		click.secho(
+			f"Bench app-cache: invalid value found for frappe version '{min_frappe}'",
+			fg="yellow",
+		)
 		# Invalid expression
 		return False
 
@@ -591,6 +595,10 @@ def remove_unused_node_modules(app_path: Path) -> None:
 			can_delete = "vite build" in build_script
 
 		if can_delete:
+			click.secho(
+				f"Bench app-cache: removing {node_modules.as_posix()}",
+				fg="yellow",
+			)
 			shutil.rmtree(node_modules)
 
 
@@ -1036,3 +1044,58 @@ def get_apps_json(path):
 
 	with open(path) as f:
 		return json.load(f)
+
+
+def is_cache_hash_valid(cache_path: Path) -> bool:
+	parts = cache_path.name.split(".")
+	if len(parts) < 2 or not parts[-2].startswith("md5-"):
+		return False
+
+	md5 = parts[-2].split("-")[1]
+	return get_file_md5(cache_path) == md5
+
+
+def unlink_no_throw(path: Path):
+	if not path.exists():
+		return
+
+	try:
+		path.unlink(True)
+	except Exception:
+		pass
+
+
+def get_app_cache_path(app_name: str, cache_key: str) -> "Optional[Path]":
+	cache_path = get_bench_cache_path("apps")
+	glob_pattern = f"{app_name}.{cache_key}.md5-*"
+
+	for app_cache_path in cache_path.glob(glob_pattern):
+		return app_cache_path
+
+	return None
+
+
+def validate_cache_and_get_path(app_name: str, cache_key: str) -> "Optional[Path]":
+	if not cache_key:
+		return
+
+	if not (cache_path := get_app_cache_path(app_name, cache_key)):
+		return
+
+	if not cache_path.is_file():
+		click.secho(
+			f"Bench app-cache: file check failed for {cache_path.as_posix()}, skipping cache",
+			fg="yellow",
+		)
+		unlink_no_throw(cache_path)
+		return
+
+	if not is_cache_hash_valid(cache_path):
+		click.secho(
+			f"Bench app-cache: hash validation failed for {cache_path.as_posix()}, skipping cache",
+			fg="yellow",
+		)
+		unlink_no_throw(cache_path)
+		return
+
+	return cache_path

bench/utils/__init__.py

@@ -5,6 +5,7 @@ import os
 import re
 import subprocess
 import sys
+import hashlib
 from functools import lru_cache
 from glob import glob
 from pathlib import Path
@@ -23,6 +24,12 @@ from bench.exceptions import (
 	InvalidRemoteException,
 )
 
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+	from typing import Optional
+
+
 logger = logging.getLogger(PROJECT_NAME)
 paths_in_app = ("hooks.py", "modules.txt", "patches.txt")
 paths_in_bench = ("apps", "sites", "config", "logs", "config/pids")
@@ -52,6 +59,7 @@ def is_frappe_app(directory: str) -> bool:
 
 	return bool(is_frappe_app)
 
+
 def get_bench_cache_path(sub_dir: Optional[str]) -> Path:
 	relative_path = "~/.cache/bench"
 	if sub_dir and not sub_dir.startswith("/"):
@@ -62,6 +70,7 @@ def get_bench_cache_path(sub_dir: Optional[str]) -> Path:
 	cache_path.mkdir(parents=True, exist_ok=True)
 	return cache_path
 
+
 @lru_cache(maxsize=None)
 def is_valid_frappe_branch(frappe_path: str, frappe_branch: str):
 	"""Check if a branch exists in a repo. Throws InvalidRemoteException if branch is not found
@@ -417,7 +426,7 @@ def get_env_frappe_commands(bench_path=".") -> List:
 	return []
 
 
-def find_org(org_repo, using_cached: bool=False):
+def find_org(org_repo, using_cached: bool = False):
 	import requests
 
 	org_repo = org_repo[0]
@@ -432,10 +441,14 @@ def find_org(org_repo, using_cached: bool=False):
 	if using_cached:
 		return "", org_repo
 
-	raise InvalidRemoteException(f"{org_repo} not found under frappe or erpnext GitHub accounts")
+	raise InvalidRemoteException(
+		f"{org_repo} not found under frappe or erpnext GitHub accounts"
+	)
 
 
-def fetch_details_from_tag(_tag: str, using_cached: bool=False) -> Tuple[str, str, str]:
+def fetch_details_from_tag(
+	_tag: str, using_cached: bool = False
+) -> Tuple[str, str, str]:
 	if not _tag:
 		raise Exception("Tag is not provided")
 
@@ -578,14 +591,17 @@ def get_cmd_from_sysargv():
 def get_app_cache_extract_filter(
 	count_threshold: int = 10_000,
 	size_threshold: int = 1_000_000_000,
-): # -> Callable[[TarInfo, str], TarInfo | None]
+):  # -> Callable[[TarInfo, str], TarInfo | None]
 	state = dict(count=0, size=0)
 
 	AbsoluteLinkError = Exception
-	def data_filter(m: TarInfo, _:str) -> TarInfo:
+
+	def data_filter(m: TarInfo, _: str) -> TarInfo:
 		return m
 
-	if (sys.version_info.major == 3 and sys.version_info.minor > 7) or sys.version_info.major > 3:
+	if (
+		sys.version_info.major == 3 and sys.version_info.minor > 7
+	) or sys.version_info.major > 3:
 		from tarfile import data_filter, AbsoluteLinkError
 
 	def filter_function(member: TarInfo, dest_path: str) -> Optional[TarInfo]:
@@ -605,3 +621,18 @@ def get_app_cache_extract_filter(
 			return None
 
 	return filter_function
+
+
+def get_file_md5(p: Path) -> "str":
+	with open(p.as_posix(), "rb") as f:
+		try:
+			file_md5 = hashlib.md5(usedforsecurity=False)
+
+		# Will throw if < 3.9, can be removed once support
+		# is dropped
+		except TypeError:
+			file_md5 = hashlib.md5()
+
+		while chunk := f.read(2**16):
+			file_md5.update(chunk)
+	return file_md5.hexdigest()

bench/utils/bench.py

@@ -688,7 +688,7 @@ def cache_list() -> None:
 		created = datetime.fromtimestamp(stat.st_ctime)
 		accessed = datetime.fromtimestamp(stat.st_atime)
 
-		app = item.name.split("-")[0]
+		app = item.name.split(".")[0]
 		tot_items += 1
 		tot_size += stat.st_size
 		compressed = item.suffix == ".tgz"
@@ -696,7 +696,7 @@ def cache_list() -> None:
 		if not printed_header:
 			click.echo(
 				f"{'APP':15}  "
-				f"{'FILE':25}  "
+				f"{'FILE':90}  "
 				f"{'SIZE':>13}  "
 				f"{'COMPRESSED'}  "
 				f"{'CREATED':19}  "
@@ -706,7 +706,7 @@ def cache_list() -> None:
 
 		click.echo(
 			f"{app:15}  "
-			f"{item.name:25}  "
+			f"{item.name:90}  "
 			f"{size_mb:10.3f} MB  "
 			f"{str(compressed):10}  "
 			f"{created:%Y-%m-%d %H:%M:%S}  "

pyproject.toml

@@ -3,7 +3,7 @@ name = "frappe-bench"
 description = "CLI to manage Multi-tenant deployments for Frappe apps"
 readme = "README.md"
 license = "GPL-3.0-only"
-requires-python = ">=3.7"
+requires-python = ">=3.8"
 authors = [
     { name = "Frappe Technologies Pvt Ltd", email = "developers@frappe.io" },
 ]