Merge pull request #98304 from jtojnar/updateScript-commit3

maintainers/scripts/update.nix: Add support for auto-commiting changes

doc/stdenv/stdenv.xml

@@ -475,9 +475,12 @@ passthru.updateScript = writeScript "update-zoom-us" ''
 <programlisting>
 passthru.updateScript = [ ../../update.sh pname "--requested-release=unstable" ];
 </programlisting>
-     </para>
-     <para>
-      The script will be usually run from the root of the Nixpkgs repository but you should not rely on that. Also note that the update scripts will be run in parallel by default; you should avoid running <command>git commit</command> or any other commands that cannot handle that.
+      The script will be run with <variable>UPDATE_NIX_ATTR_PATH</variable> environment variable set to the attribute path it is supposed to update.
+      <note>
+       <para>
+        The script will be usually run from the root of the Nixpkgs repository but you should not rely on that. Also note that the update scripts will be run in parallel by default; you should avoid running <command>git commit</command> or any other commands that cannot handle that.
+       </para>
+      </note>
      </para>
      <para>
       For information about how to run the updates, execute <command>nix-shell maintainers/scripts/update.nix</command>.

maintainers/scripts/update.nix

@@ -4,6 +4,7 @@
 , max-workers ? null
 , include-overlays ? false
 , keep-going ? null
+, commit ? null
 }:
 
 # TODO: add assert statements
@@ -31,30 +32,47 @@ let
       in
         [x] ++ nubOn f xs;
 
-  packagesWithPath = relativePath: cond: return: pathContent:
+  /* Recursively find all packages (derivations) in `pkgs` matching `cond` predicate.
+
+    Type: packagesWithPath :: AttrPath → (AttrPath → derivation → bool) → (AttrSet | List) → List<AttrSet{attrPath :: str; package :: derivation; }>
+          AttrPath :: [str]
+
+    The packages will be returned as a list of named pairs comprising of:
+      - attrPath: stringified attribute path (based on `rootPath`)
+      - package: corresponding derivation
+   */
+  packagesWithPath = rootPath: cond: pkgs:
     let
-      result = builtins.tryEval pathContent;
-
-      dedupResults = lst: nubOn (pkg: pkg.updateScript) (lib.concatLists lst);
-    in
-      if result.success then
+      packagesWithPathInner = path: pathContent:
         let
-          pathContent = result.value;
-        in
-          if lib.isDerivation pathContent then
-            lib.optional (cond relativePath pathContent) (return relativePath pathContent)
-          else if lib.isAttrs pathContent then
-            # If user explicitly points to an attrSet or it is marked for recursion, we recur.
-            if relativePath == [] || pathContent.recurseForDerivations or false || pathContent.recurseForRelease or false then
-              dedupResults (lib.mapAttrsToList (name: elem: packagesWithPath (relativePath ++ [name]) cond return elem) pathContent)
-            else []
-          else if lib.isList pathContent then
-            dedupResults (lib.imap0 (i: elem: packagesWithPath (relativePath ++ [i]) cond return elem) pathContent)
-          else []
-      else [];
+          result = builtins.tryEval pathContent;
 
+          dedupResults = lst: nubOn ({ package, attrPath }: package.updateScript) (lib.concatLists lst);
+        in
+          if result.success then
+            let
+              evaluatedPathContent = result.value;
+            in
+              if lib.isDerivation evaluatedPathContent then
+                lib.optional (cond path evaluatedPathContent) { attrPath = lib.concatStringsSep "." path; package = evaluatedPathContent; }
+              else if lib.isAttrs evaluatedPathContent then
+                # If user explicitly points to an attrSet or it is marked for recursion, we recur.
+                if path == rootPath || evaluatedPathContent.recurseForDerivations or false || evaluatedPathContent.recurseForRelease or false then
+                  dedupResults (lib.mapAttrsToList (name: elem: packagesWithPathInner (path ++ [name]) elem) evaluatedPathContent)
+                else []
+              else if lib.isList evaluatedPathContent then
+                dedupResults (lib.imap0 (i: elem: packagesWithPathInner (path ++ [i]) elem) evaluatedPathContent)
+              else []
+          else [];
+    in
+      packagesWithPathInner rootPath pkgs;
+
+  /* Recursively find all packages (derivations) in `pkgs` matching `cond` predicate.
+   */
   packagesWith = packagesWithPath [];
 
+  /* Recursively find all packages in `pkgs` with updateScript by given maintainer.
+   */
   packagesWithUpdateScriptAndMaintainer = maintainer':
     let
       maintainer =
@@ -63,47 +81,51 @@ let
         else
           builtins.getAttr maintainer' lib.maintainers;
     in
-      packagesWith (relativePath: pkg: builtins.hasAttr "updateScript" pkg &&
-                                 (if builtins.hasAttr "maintainers" pkg.meta
-                                   then (if builtins.isList pkg.meta.maintainers
-                                           then builtins.elem maintainer pkg.meta.maintainers
-                                           else maintainer == pkg.meta.maintainers
-                                        )
-                                   else false
-                                 )
-                   )
-                   (relativePath: pkg: pkg)
-                   pkgs;
+      packagesWith (path: pkg: builtins.hasAttr "updateScript" pkg &&
+                         (if builtins.hasAttr "maintainers" pkg.meta
+                           then (if builtins.isList pkg.meta.maintainers
+                                   then builtins.elem maintainer pkg.meta.maintainers
+                                   else maintainer == pkg.meta.maintainers
+                                )
+                           else false
+                         )
+                   );
 
-  packagesWithUpdateScript = path:
+  /* Recursively find all packages under `path` in `pkgs` with updateScript.
+   */
+  packagesWithUpdateScript = path: pkgs:
     let
-      pathContent = lib.attrByPath (lib.splitString "." path) null pkgs;
+      prefix = lib.splitString "." path;
+      pathContent = lib.attrByPath prefix null pkgs;
     in
       if pathContent == null then
         builtins.throw "Attribute path `${path}` does not exists."
       else
-        packagesWith (relativePath: pkg: builtins.hasAttr "updateScript" pkg)
-                       (relativePath: pkg: pkg)
+        packagesWithPath prefix (path: pkg: builtins.hasAttr "updateScript" pkg)
                        pathContent;
 
-  packageByName = name:
+  /* Find a package under `path` in `pkgs` and require that it has an updateScript.
+   */
+  packageByName = path: pkgs:
     let
-        package = lib.attrByPath (lib.splitString "." name) null pkgs;
+        package = lib.attrByPath (lib.splitString "." path) null pkgs;
     in
       if package == null then
-        builtins.throw "Package with an attribute name `${name}` does not exists."
+        builtins.throw "Package with an attribute name `${path}` does not exists."
       else if ! builtins.hasAttr "updateScript" package then
-        builtins.throw "Package with an attribute name `${name}` does not have a `passthru.updateScript` attribute defined."
+        builtins.throw "Package with an attribute name `${path}` does not have a `passthru.updateScript` attribute defined."
       else
-        package;
+        { attrPath = path; inherit package; };
 
+  /* List of packages matched based on the CLI arguments.
+   */
   packages =
     if package != null then
-      [ (packageByName package) ]
+      [ (packageByName package pkgs) ]
     else if maintainer != null then
-      packagesWithUpdateScriptAndMaintainer maintainer
+      packagesWithUpdateScriptAndMaintainer maintainer pkgs
     else if path != null then
-      packagesWithUpdateScript path
+      packagesWithUpdateScript path pkgs
     else
       builtins.throw "No arguments provided.\n\n${helpText}";
 
@@ -132,19 +154,32 @@ let
         --argstr keep-going true
 
     to continue running when a single update fails.
+
+    You can also make the updater automatically commit on your behalf from updateScripts
+    that support it by adding
+
+        --argstr commit true
   '';
 
-  packageData = package: {
+  /* Transform a matched package into an object for update.py.
+   */
+  packageData = { package, attrPath }: {
     name = package.name;
     pname = lib.getName package;
-    updateScript = map builtins.toString (lib.toList package.updateScript);
+    oldVersion = lib.getVersion package;
+    updateScript = map builtins.toString (lib.toList (package.updateScript.command or package.updateScript));
+    supportedFeatures = package.updateScript.supportedFeatures or [];
+    attrPath = package.updateScript.attrPath or attrPath;
   };
 
+  /* JSON file with data for update.py.
+   */
   packagesJson = pkgs.writeText "packages.json" (builtins.toJSON (map packageData packages));
 
   optionalArgs =
     lib.optional (max-workers != null) "--max-workers=${max-workers}"
-    ++ lib.optional (keep-going == "true") "--keep-going";
+    ++ lib.optional (keep-going == "true") "--keep-going"
+    ++ lib.optional (commit == "true") "--commit";
 
   args = [ packagesJson ] ++ optionalArgs;
 

maintainers/scripts/update.py

@@ -1,23 +1,189 @@
+from __future__ import annotations
+from typing import Dict, Generator, List, Optional, Tuple
 import argparse
-import concurrent.futures
+import asyncio
+import contextlib
 import json
 import os
+import re
 import subprocess
 import sys
+import tempfile
 
-updates = {}
+class CalledProcessError(Exception):
+    process: asyncio.subprocess.Process
 
 def eprint(*args, **kwargs):
     print(*args, file=sys.stderr, **kwargs)
 
-def run_update_script(package):
+async def check_subprocess(*args, **kwargs):
+    """
+    Emulate check argument of subprocess.run function.
+    """
+    process = await asyncio.create_subprocess_exec(*args, **kwargs)
+    returncode = await process.wait()
+
+    if returncode != 0:
+        error = CalledProcessError()
+        error.process = process
+
+        raise error
+
+    return process
+
+async def run_update_script(nixpkgs_root: str, merge_lock: asyncio.Lock, temp_dir: Optional[Tuple[str, str]], package: Dict, keep_going: bool):
+    worktree: Optional[str] = None
+
+    update_script_command = package['updateScript']
+
+    if temp_dir is not None:
+        worktree, _branch = temp_dir
+
+        # Update scripts can use $(dirname $0) to get their location but we want to run
+        # their clones in the git worktree, not in the main nixpkgs repo.
+        update_script_command = map(lambda arg: re.sub(r'^{0}'.format(re.escape(nixpkgs_root)), worktree, arg), update_script_command)
+
     eprint(f" - {package['name']}: UPDATING ...")
 
-    subprocess.run(package['updateScript'], stdout=subprocess.PIPE, stderr=subprocess.STDOUT, check=True)
+    try:
+        update_process = await check_subprocess('env', f"UPDATE_NIX_ATTR_PATH={package['attrPath']}", *update_script_command, stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.PIPE, cwd=worktree)
+        update_info = await update_process.stdout.read()
 
+        await merge_changes(merge_lock, package, update_info, temp_dir)
+    except KeyboardInterrupt as e:
+        eprint('Cancelling…')
+        raise asyncio.exceptions.CancelledError()
+    except CalledProcessError as e:
+        eprint(f" - {package['name']}: ERROR")
+        eprint()
+        eprint(f"--- SHOWING ERROR LOG FOR {package['name']} ----------------------")
+        eprint()
+        stderr = await e.process.stderr.read()
+        eprint(stderr.decode('utf-8'))
+        with open(f"{package['pname']}.log", 'wb') as logfile:
+            logfile.write(stderr)
+        eprint()
+        eprint(f"--- SHOWING ERROR LOG FOR {package['name']} ----------------------")
 
-def main(max_workers, keep_going, packages):
-    with open(sys.argv[1]) as f:
+        if not keep_going:
+            raise asyncio.exceptions.CancelledError()
+
+@contextlib.contextmanager
+def make_worktree() -> Generator[Tuple[str, str], None, None]:
+    with tempfile.TemporaryDirectory() as wt:
+        branch_name = f'update-{os.path.basename(wt)}'
+        target_directory = f'{wt}/nixpkgs'
+
+        subprocess.run(['git', 'worktree', 'add', '-b', branch_name, target_directory])
+        yield (target_directory, branch_name)
+        subprocess.run(['git', 'worktree', 'remove', '--force', target_directory])
+        subprocess.run(['git', 'branch', '-D', branch_name])
+
+async def commit_changes(name: str, merge_lock: asyncio.Lock, worktree: str, branch: str, changes: List[Dict]) -> None:
+    for change in changes:
+        # Git can only handle a single index operation at a time
+        async with merge_lock:
+            await check_subprocess('git', 'add', *change['files'], cwd=worktree)
+            commit_message = '{attrPath}: {oldVersion} → {newVersion}'.format(**change)
+            await check_subprocess('git', 'commit', '--quiet', '-m', commit_message, cwd=worktree)
+            await check_subprocess('git', 'cherry-pick', branch)
+
+async def check_changes(package: Dict, worktree: str, update_info: str):
+    if 'commit' in package['supportedFeatures']:
+        changes = json.loads(update_info)
+    else:
+        changes = [{}]
+
+    # Try to fill in missing attributes when there is just a single change.
+    if len(changes) == 1:
+        # Dynamic data from updater take precedence over static data from passthru.updateScript.
+        if 'attrPath' not in changes[0]:
+            # update.nix is always passing attrPath
+            changes[0]['attrPath'] = package['attrPath']
+
+        if 'oldVersion' not in changes[0]:
+            # update.nix is always passing oldVersion
+            changes[0]['oldVersion'] = package['oldVersion']
+
+        if 'newVersion' not in changes[0]:
+            attr_path = changes[0]['attrPath']
+            obtain_new_version_process = await check_subprocess('nix-instantiate', '--expr', f'with import ./. {{}}; lib.getVersion {attr_path}', '--eval', '--strict', '--json', stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.PIPE, cwd=worktree)
+            changes[0]['newVersion'] = json.loads((await obtain_new_version_process.stdout.read()).decode('utf-8'))
+
+        if 'files' not in changes[0]:
+            changed_files_process = await check_subprocess('git', 'diff', '--name-only', stdout=asyncio.subprocess.PIPE, cwd=worktree)
+            changed_files = (await changed_files_process.stdout.read()).splitlines()
+            changes[0]['files'] = changed_files
+
+            if len(changed_files) == 0:
+                return []
+
+    return changes
+
+async def merge_changes(merge_lock: asyncio.Lock, package: Dict, update_info: str, temp_dir: Optional[Tuple[str, str]]) -> None:
+    if temp_dir is not None:
+        worktree, branch = temp_dir
+        changes = await check_changes(package, worktree, update_info)
+
+        if len(changes) > 0:
+            await commit_changes(package['name'], merge_lock, worktree, branch, changes)
+        else:
+            eprint(f" - {package['name']}: DONE, no changes.")
+    else:
+        eprint(f" - {package['name']}: DONE.")
+
+async def updater(nixpkgs_root: str, temp_dir: Optional[Tuple[str, str]], merge_lock: asyncio.Lock, packages_to_update: asyncio.Queue[Optional[Dict]], keep_going: bool, commit: bool):
+    while True:
+        package = await packages_to_update.get()
+        if package is None:
+            # A sentinel received, we are done.
+            return
+
+        if not ('commit' in package['supportedFeatures'] or 'attrPath' in package):
+            temp_dir = None
+
+        await run_update_script(nixpkgs_root, merge_lock, temp_dir, package, keep_going)
+
+async def start_updates(max_workers: int, keep_going: bool, commit: bool, packages: List[Dict]):
+    merge_lock = asyncio.Lock()
+    packages_to_update: asyncio.Queue[Optional[Dict]] = asyncio.Queue()
+
+    with contextlib.ExitStack() as stack:
+        temp_dirs: List[Optional[Tuple[str, str]]] = []
+
+        # Do not create more workers than there are packages.
+        num_workers = min(max_workers, len(packages))
+
+        nixpkgs_root_process = await check_subprocess('git', 'rev-parse', '--show-toplevel', stdout=asyncio.subprocess.PIPE)
+        nixpkgs_root = (await nixpkgs_root_process.stdout.read()).decode('utf-8').strip()
+
+        # Set up temporary directories when using auto-commit.
+        for i in range(num_workers):
+            temp_dir = stack.enter_context(make_worktree()) if commit else None
+            temp_dirs.append(temp_dir)
+
+        # Fill up an update queue,
+        for package in packages:
+            await packages_to_update.put(package)
+
+        # Add sentinels, one for each worker.
+        # A workers will terminate when it gets sentinel from the queue.
+        for i in range(num_workers):
+            await packages_to_update.put(None)
+
+        # Prepare updater workers for each temp_dir directory.
+        # At most `num_workers` instances of `run_update_script` will be running at one time.
+        updaters = asyncio.gather(*[updater(nixpkgs_root, temp_dir, merge_lock, packages_to_update, keep_going, commit) for temp_dir in temp_dirs])
+
+        try:
+            # Start updater workers.
+            await updaters
+        except asyncio.exceptions.CancelledError as e:
+            # When one worker is cancelled, cancel the others too.
+            updaters.cancel()
+
+def main(max_workers: int, keep_going: bool, commit: bool, packages_path: str) -> None:
+    with open(packages_path) as f:
         packages = json.load(f)
 
     eprint()
@@ -31,29 +197,7 @@ def main(max_workers, keep_going, packages):
         eprint()
         eprint('Running update for:')
 
-        with concurrent.futures.ProcessPoolExecutor(max_workers=max_workers) as executor:
-            for package in packages:
-                updates[executor.submit(run_update_script, package)] = package
-
-            for future in concurrent.futures.as_completed(updates):
-                package = updates[future]
-
-                try:
-                    future.result()
-                    eprint(f" - {package['name']}: DONE.")
-                except subprocess.CalledProcessError as e:
-                    eprint(f" - {package['name']}: ERROR")
-                    eprint()
-                    eprint(f"--- SHOWING ERROR LOG FOR {package['name']} ----------------------")
-                    eprint()
-                    eprint(e.stdout.decode('utf-8'))
-                    with open(f"{package['pname']}.log", 'wb') as f:
-                        f.write(e.stdout)
-                    eprint()
-                    eprint(f"--- SHOWING ERROR LOG FOR {package['name']} ----------------------")
-
-                    if not keep_going:
-                        sys.exit(1)
+        asyncio.run(start_updates(max_workers, keep_going, commit, packages))
 
         eprint()
         eprint('Packages updated!')
@@ -65,15 +209,14 @@ def main(max_workers, keep_going, packages):
 parser = argparse.ArgumentParser(description='Update packages')
 parser.add_argument('--max-workers', '-j', dest='max_workers', type=int, help='Number of updates to run concurrently', nargs='?', default=4)
 parser.add_argument('--keep-going', '-k', dest='keep_going', action='store_true', help='Do not stop after first failure')
+parser.add_argument('--commit', '-c', dest='commit', action='store_true', help='Commit the changes')
 parser.add_argument('packages', help='JSON file containing the list of package names and their update scripts')
 
 if __name__ == '__main__':
     args = parser.parse_args()
 
     try:
-        main(args.max_workers, args.keep_going, args.packages)
-    except (KeyboardInterrupt, SystemExit) as e:
-        for update in updates:
-            update.cancel()
-
-        sys.exit(e.code if isinstance(e, SystemExit) else 130)
+        main(args.max_workers, args.keep_going, args.commit, args.packages)
+    except KeyboardInterrupt as e:
+        # Let’s cancel outside of the main loop too.
+        sys.exit(130)

pkgs/common-updater/scripts/update-source-version

@@ -11,7 +11,7 @@ die() {
 usage() {
     echo "Usage: $scriptName <attr> <version> [<new-source-hash>] [<new-source-url>]"
     echo "                              [--version-key=<version-key>] [--system=<system>] [--file=<file-to-update>]"
-    echo "                              [--ignore-same-hash]"
+    echo "                              [--ignore-same-hash] [--print-changes]"
 }
 
 args=()
@@ -33,6 +33,9 @@ for arg in "$@"; do
         --ignore-same-hash)
             ignoreSameHash="true"
         ;;
+        --print-changes)
+            printChanges="true"
+        ;;
         --help)
             usage
             exit 0
@@ -102,6 +105,9 @@ fi
 
 if [[ "$oldVersion" = "$newVersion" ]]; then
     echo "$scriptName: New version same as old version, nothing to do." >&2
+    if [ -n "$printChanges" ]; then
+        printf '[]\n'
+    fi
     exit 0
 fi
 
@@ -197,3 +203,7 @@ fi
 
 rm -f "$nixFile.bak"
 rm -f "$attr.fetchlog"
+
+if [ -n "$printChanges" ]; then
+    printf '[{"attrPath":"%s","oldVersion":"%s","newVersion":"%s","files":["%s"]}]\n' "$attr" "$oldVersion" "$newVersion" "$nixFile"
+fi