core: code changes to handle branch-permissions / locking / incorrect client support

vcsserver/exceptions.py

@@ -1,131 +1,138 b''
 # RhodeCode VCSServer provides access to different vcs backends via network.
 # Copyright (C) 2014-2023 RhodeCode GmbH
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
 # the Free Software Foundation; either version 3 of the License, or
 # (at your option) any later version.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
 # along with this program; if not, write to the Free Software Foundation,
 # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301  USA
 
 """
 Special exception handling over the wire.
 
 Since we cannot assume that our client is able to import our exception classes,
 this module provides a "wrapping" mechanism to raise plain exceptions
 which contain an extra attribute `_vcs_kind` to allow a client to distinguish
 different error conditions.
 """
 
 from pyramid.httpexceptions import HTTPLocked, HTTPForbidden
 
 
 def _make_exception(kind, org_exc, *args):
     """
     Prepares a base `Exception` instance to be sent over the wire.
 
     To give our caller a hint what this is about, it will attach an attribute
     `_vcs_kind` to the exception.
     """
     exc = Exception(*args)
     exc._vcs_kind = kind
     exc._org_exc = org_exc
     exc._org_exc_tb = getattr(org_exc, '_org_exc_tb', '')
     return exc
 
 
 def AbortException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('abort', org_exc, *args)
     return _make_exception_wrapper
 
 
 def ArchiveException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('archive', org_exc, *args)
     return _make_exception_wrapper
 
 
-def ClientNotSupportedException(org_exc=None):
-    def _make_exception_wrapper(*args):
-        return _make_exception('client_not_supported', org_exc, *args)
-    return _make_exception_wrapper
-
-
 def LookupException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('lookup', org_exc, *args)
     return _make_exception_wrapper
 
 
 def VcsException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('error', org_exc, *args)
     return _make_exception_wrapper
 
 
 def LockedRepoException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('repo_locked', org_exc, *args)
     return _make_exception_wrapper
 
 
 def RepositoryBranchProtectedException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('repo_branch_protected', org_exc, *args)
     return _make_exception_wrapper
 
+def ClientNotSupportedException(org_exc=None):
+    def _make_exception_wrapper(*args):
+        return _make_exception('client_not_supported', org_exc, *args)
+    return _make_exception_wrapper
 
 def RequirementException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('requirement', org_exc, *args)
     return _make_exception_wrapper
 
 
 def UnhandledException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('unhandled', org_exc, *args)
     return _make_exception_wrapper
 
 
 def URLError(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('url_error', org_exc, *args)
     return _make_exception_wrapper
 
 
 def SubrepoMergeException(org_exc=None):
     def _make_exception_wrapper(*args):
         return _make_exception('subrepo_merge_error', org_exc, *args)
     return _make_exception_wrapper
 
 
 class HTTPRepoLocked(HTTPLocked):
     """
     Subclass of HTTPLocked response that allows to set the title and status
     code via constructor arguments.
     """
     def __init__(self, title, status_code=None, **kwargs):
         self.code = status_code or HTTPLocked.code
         self.title = title
         super().__init__(**kwargs)
 
 
-class HTTPRepoBranchProtected(HTTPForbidden):
-    def __init__(self, *args, **kwargs):
-        super(HTTPForbidden, self).__init__(*args, **kwargs)
+class HTTPRepoBranchProtected(HTTPLocked):
+    def __init__(self, title, status_code=None, **kwargs):
+        self.code = status_code or HTTPLocked.code
+        self.title = title
+        super().__init__(**kwargs)
+
+
+class HTTPClientNotSupported(HTTPLocked):
+    def __init__(self, title, status_code=None, **kwargs):
+        self.code = status_code or HTTPLocked.code
+        self.title = title
+        super().__init__(**kwargs)
 
 
 class RefNotFoundException(KeyError):
     pass
 
 
 class NoContentException(ValueError):
     pass

vcsserver/hooks.py

@@ -1,780 +1,782 b''
 # RhodeCode VCSServer provides access to different vcs backends via network.
-# Copyright (C) 2014-2023 RhodeCode GmbH
+# Copyright (C) 2014-2024 RhodeCode GmbH
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
 # the Free Software Foundation; either version 3 of the License, or
 # (at your option) any later version.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
 # along with this program; if not, write to the Free Software Foundation,
 # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301  USA
 
 import io
 import os
 import sys
 import logging
 import collections
 import base64
 import msgpack
 import dataclasses
 import pygit2
 
 import http.client
 from celery import Celery
 
 import mercurial.scmutil
 import mercurial.node
 
 from vcsserver import exceptions, subprocessio, settings
 from vcsserver.lib.ext_json import json
 from vcsserver.lib.str_utils import ascii_str, safe_str
 from vcsserver.lib.svn_txn_utils import get_txn_id_from_store
 from vcsserver.remote.git_remote import Repository
 
 celery_app = Celery('__vcsserver__')
 log = logging.getLogger(__name__)
 
 
 class HooksCeleryClient:
     TASK_TIMEOUT = 60  # time in seconds
 
     def __init__(self, queue, backend):
         celery_app.config_from_object({
             'broker_url': queue, 'result_backend': backend,
             'broker_connection_retry_on_startup': True,
             'task_serializer': 'json',
             'accept_content': ['json', 'msgpack'],
             'result_serializer': 'json',
             'result_accept_content': ['json', 'msgpack']
         })
         self.celery_app = celery_app
 
     def __call__(self, method, extras):
+        # NOTE: exception handling for those tasks executed is in
+        # @adapt_for_celery decorator
+        # also see: _maybe_handle_exception which is handling exceptions
         inquired_task = self.celery_app.signature(
             f'rhodecode.lib.celerylib.tasks.{method}'
         )
         result = inquired_task.delay(extras).get(timeout=self.TASK_TIMEOUT)
 
         return result
 
 
 class HooksShadowRepoClient:
 
     def __call__(self, hook_name, extras):
         return {'output': '', 'status': 0}
 
 
 class RemoteMessageWriter:
     """Writer base class."""
     def write(self, message):
         raise NotImplementedError()
 
 
 class HgMessageWriter(RemoteMessageWriter):
     """Writer that knows how to send messages to mercurial clients."""
 
     def __init__(self, ui):
         self.ui = ui
 
     def write(self, message: str):
-        # TODO: Check why the quiet flag is set by default.
-        old = self.ui.quiet
-        self.ui.quiet = False
-        self.ui.status(message.encode('utf-8'))
-        self.ui.quiet = old
+        args = (message.encode('utf-8'),)
+        self.ui._writemsg(self.ui._fmsgerr, type=b'status', *args)
 
 
 class GitMessageWriter(RemoteMessageWriter):
     """Writer that knows how to send messages to git clients."""
 
     def __init__(self, stdout=None):
         self.stdout = stdout or sys.stdout
 
     def write(self, message: str):
-        self.stdout.write(message)
+        self.stdout.write(message + "\n" if message else "")
 
 
 class SvnMessageWriter(RemoteMessageWriter):
     """Writer that knows how to send messages to svn clients."""
 
     def __init__(self, stderr=None):
         # SVN needs data sent to stderr for back-to-client messaging
         self.stderr = stderr or sys.stderr
 
     def write(self, message):
         self.stderr.write(message)
 
 
-def _maybe_handle_exception(result):
-
+def _maybe_handle_exception(writer, result):
+    """
+    adopt_for_celery defines the exception/exception_traceback
+    Ths result is a direct output from a celery task
+    """
 
     exception_class = result.get('exception')
     exception_traceback = result.get('exception_traceback')
-    if not exception_class:
-        return
 
-    log.debug('Handling hook-call exception: %s', exception_class)
-
-    if exception_traceback:
-        log.error('Got traceback from remote call:%s', exception_traceback)
-
-    if exception_class == 'HTTPLockedRepo':
-        raise exceptions.LockedRepoException()(*result['exception_args'])
-    elif exception_class == 'ClientNotSupportedError':
-        raise exceptions.ClientNotSupportedException()(*result['exception_args'])
-    elif exception_class == 'HTTPBranchProtected':
-        raise exceptions.RepositoryBranchProtectedException()(*result['exception_args'])
-    elif exception_class == 'RepositoryError':
-        raise exceptions.VcsException()(*result['exception_args'])
-    elif exception_class:
-        raise Exception(
-            f"""Got remote exception "{exception_class}" with args "{result['exception_args']}" """
-        )
+    match exception_class:
+        # NOTE: the underlying exceptions are setting _vcs_kind special marker
+        # which is later handled by `handle_vcs_exception` and translated into a special HTTP exception
+        # propagated later to the client
+        case 'HTTPLockedRepo':
+            raise exceptions.LockedRepoException()(*result['exception_args'])
+        case 'ClientNotSupported':
+            raise exceptions.ClientNotSupportedException()(*result['exception_args'])
+        case 'HTTPBranchProtected':
+            raise exceptions.RepositoryBranchProtectedException()(*result['exception_args'])
+        case 'RepositoryError':
+            raise exceptions.VcsException()(*result['exception_args'])
+        case _:
+            if exception_class:
+                log.error('Handling hook-call exception. Got traceback from remote call:%s', exception_traceback)
+                raise Exception(
+                    f"""Got remote exception "{exception_class}" with args "{result['exception_args']}" """
+                )
 
 
 def _get_hooks_client(extras):
     task_queue = extras.get('task_queue')
     task_backend = extras.get('task_backend')
     is_shadow_repo = extras.get('is_shadow_repo')
 
     if task_queue and task_backend:
         return HooksCeleryClient(task_queue, task_backend)
     elif is_shadow_repo:
         return HooksShadowRepoClient()
     else:
         raise Exception("Hooks client not found!")
 
 
 def _call_hook(hook_name, extras, writer):
     hooks_client = _get_hooks_client(extras)
     log.debug('Hooks, using client:%s', hooks_client)
     result = hooks_client(hook_name, extras)
     log.debug('Hooks got result: %s', result)
-    _maybe_handle_exception(result)
+    _maybe_handle_exception(writer, result)
     writer.write(result['output'])
 
     return result['status']
 
 
 def _extras_from_ui(ui):
     hook_data = ui.config(b'rhodecode', b'RC_SCM_DATA')
     if not hook_data:
         # maybe it's inside environ ?
         env_hook_data = os.environ.get('RC_SCM_DATA')
         if env_hook_data:
             hook_data = env_hook_data
 
     extras = {}
     if hook_data:
         extras = json.loads(hook_data)
     return extras
 
 
 def _rev_range_hash(repo, node, check_heads=False):
     from vcsserver.hgcompat import get_ctx
 
     commits = []
     revs = []
     start = get_ctx(repo, node).rev()
     end = len(repo)
     for rev in range(start, end):
         revs.append(rev)
         ctx = get_ctx(repo, rev)
         commit_id = ascii_str(mercurial.node.hex(ctx.node()))
         branch = safe_str(ctx.branch())
         commits.append((commit_id, branch))
 
     parent_heads = []
     if check_heads:
         parent_heads = _check_heads(repo, start, end, revs)
     return commits, parent_heads
 
 
 def _check_heads(repo, start, end, commits):
     from vcsserver.hgcompat import get_ctx
     changelog = repo.changelog
     parents = set()
 
     for new_rev in commits:
         for p in changelog.parentrevs(new_rev):
             if p == mercurial.node.nullrev:
                 continue
             if p < start:
                 parents.add(p)
 
     for p in parents:
         branch = get_ctx(repo, p).branch()
         # The heads descending from that parent, on the same branch
         parent_heads = {p}
         reachable = {p}
         for x in range(p + 1, end):
             if get_ctx(repo, x).branch() != branch:
                 continue
             for pp in changelog.parentrevs(x):
                 if pp in reachable:
                     reachable.add(x)
                     parent_heads.discard(pp)
                     parent_heads.add(x)
         # More than one head? Suggest merging
         if len(parent_heads) > 1:
             return list(parent_heads)
 
     return []
 
 
 def _get_git_env():
     env = {}
     for k, v in os.environ.items():
         if k.startswith('GIT'):
             env[k] = v
 
     # serialized version
     return [(k, v) for k, v in env.items()]
 
 
 def _get_hg_env(old_rev, new_rev, txnid, repo_path):
     env = {}
     for k, v in os.environ.items():
         if k.startswith('HG'):
             env[k] = v
 
     env['HG_NODE'] = old_rev
     env['HG_NODE_LAST'] = new_rev
     env['HG_TXNID'] = txnid
     env['HG_PENDING'] = repo_path
 
     return [(k, v) for k, v in env.items()]
 
 
 def _get_ini_settings(ini_file):
     from vcsserver.http_main import sanitize_settings_and_apply_defaults
     from vcsserver.lib.config_utils import get_app_config_lightweight, configure_and_store_settings
 
     global_config = {'__file__': ini_file}
     ini_settings = get_app_config_lightweight(ini_file)
     sanitize_settings_and_apply_defaults(global_config, ini_settings)
     configure_and_store_settings(global_config, ini_settings)
 
     return ini_settings
 
 
 def _fix_hooks_executables(ini_path=''):
     """
     This is a trick to set proper settings.EXECUTABLE paths for certain execution patterns
     especially for subversion where hooks strip entire env, and calling just 'svn' command will most likely fail
     because svn is not on PATH
     """
     # set defaults, in case we can't read from ini_file
     core_binary_dir = settings.BINARY_DIR or '/usr/local/bin/rhodecode_bin/vcs_bin'
     if ini_path:
         ini_settings = _get_ini_settings(ini_path)
         core_binary_dir = ini_settings['core.binary_dir']
 
     settings.BINARY_DIR = core_binary_dir
 
 
 def repo_size(ui, repo, **kwargs):
     extras = _extras_from_ui(ui)
     return _call_hook('repo_size', extras, HgMessageWriter(ui))
 
 
 def pre_pull(ui, repo, **kwargs):
     extras = _extras_from_ui(ui)
     return _call_hook('pre_pull', extras, HgMessageWriter(ui))
 
 
 def pre_pull_ssh(ui, repo, **kwargs):
     extras = _extras_from_ui(ui)
     if extras and extras.get('SSH'):
         return pre_pull(ui, repo, **kwargs)
     return 0
 
 
 def post_pull(ui, repo, **kwargs):
     extras = _extras_from_ui(ui)
     return _call_hook('post_pull', extras, HgMessageWriter(ui))
 
 
 def post_pull_ssh(ui, repo, **kwargs):
     extras = _extras_from_ui(ui)
     if extras and extras.get('SSH'):
         return post_pull(ui, repo, **kwargs)
     return 0
 
 
 def pre_push(ui, repo, node=None, **kwargs):
     """
     Mercurial pre_push hook
     """
     extras = _extras_from_ui(ui)
     detect_force_push = extras.get('detect_force_push')
 
     rev_data = []
     hook_type: str = safe_str(kwargs.get('hooktype'))
 
     if node and hook_type == 'pretxnchangegroup':
         branches = collections.defaultdict(list)
         commits, _heads = _rev_range_hash(repo, node, check_heads=detect_force_push)
         for commit_id, branch in commits:
             branches[branch].append(commit_id)
 
         for branch, commits in branches.items():
             old_rev = ascii_str(kwargs.get('node_last')) or commits[0]
             rev_data.append({
                 'total_commits': len(commits),
                 'old_rev': old_rev,
                 'new_rev': commits[-1],
                 'ref': '',
                 'type': 'branch',
                 'name': branch,
             })
 
         for push_ref in rev_data:
             push_ref['multiple_heads'] = _heads
 
             repo_path = os.path.join(
                 extras.get('repo_store', ''), extras.get('repository', ''))
             push_ref['hg_env'] = _get_hg_env(
                 old_rev=push_ref['old_rev'],
                 new_rev=push_ref['new_rev'], txnid=ascii_str(kwargs.get('txnid')),
                 repo_path=repo_path)
 
     extras['hook_type'] = hook_type or 'pre_push'
     extras['commit_ids'] = rev_data
 
     return _call_hook('pre_push', extras, HgMessageWriter(ui))
 
 
 def pre_push_ssh(ui, repo, node=None, **kwargs):
     extras = _extras_from_ui(ui)
     if extras.get('SSH'):
         return pre_push(ui, repo, node, **kwargs)
 
     return 0
 
 
 def pre_push_ssh_auth(ui, repo, node=None, **kwargs):
     """
     Mercurial pre_push hook for SSH
     """
     extras = _extras_from_ui(ui)
     if extras.get('SSH'):
         permission = extras['SSH_PERMISSIONS']
 
         if 'repository.write' == permission or 'repository.admin' == permission:
             return 0
 
         # non-zero ret code
         return 1
 
     return 0
 
 
 def post_push(ui, repo, node, **kwargs):
     """
     Mercurial post_push hook
     """
     extras = _extras_from_ui(ui)
 
     commit_ids = []
     branches = []
     bookmarks = []
     tags = []
     hook_type: str = safe_str(kwargs.get('hooktype'))
 
     commits, _heads = _rev_range_hash(repo, node)
     for commit_id, branch in commits:
         commit_ids.append(commit_id)
         if branch not in branches:
             branches.append(branch)
 
     if hasattr(ui, '_rc_pushkey_bookmarks'):
         bookmarks = ui._rc_pushkey_bookmarks
 
     extras['hook_type'] = hook_type or 'post_push'
     extras['commit_ids'] = commit_ids
 
     extras['new_refs'] = {
         'branches': branches,
         'bookmarks': bookmarks,
         'tags': tags
     }
 
     return _call_hook('post_push', extras, HgMessageWriter(ui))
 
 
 def post_push_ssh(ui, repo, node, **kwargs):
     """
     Mercurial post_push hook for SSH
     """
     if _extras_from_ui(ui).get('SSH'):
         return post_push(ui, repo, node, **kwargs)
     return 0
 
 
 def key_push(ui, repo, **kwargs):
     from vcsserver.hgcompat import get_ctx
 
     if kwargs['new'] != b'0' and kwargs['namespace'] == b'bookmarks':
         # store new bookmarks in our UI object propagated later to post_push
         ui._rc_pushkey_bookmarks = get_ctx(repo, kwargs['key']).bookmarks()
     return
 
 
 # backward compat
 log_pull_action = post_pull
 
 # backward compat
 log_push_action = post_push
 
 
 def handle_git_pre_receive(unused_repo_path, unused_revs, unused_env):
     """
     Old hook name: keep here for backward compatibility.
 
     This is only required when the installed git hooks are not upgraded.
     """
     pass
 
 
 def handle_git_post_receive(unused_repo_path, unused_revs, unused_env):
     """
     Old hook name: keep here for backward compatibility.
 
     This is only required when the installed git hooks are not upgraded.
     """
     pass
 
 
 @dataclasses.dataclass
 class HookResponse:
     status: int
     output: str
 
 
 def git_pre_pull(extras) -> HookResponse:
     """
     Pre pull hook.
 
     :param extras: dictionary containing the keys defined in simplevcs
     :type extras: dict
 
     :return: status code of the hook. 0 for success.
     :rtype: int
     """
 
     if 'pull' not in extras['hooks']:
         return HookResponse(0, '')
 
     stdout = io.StringIO()
     try:
         status_code = _call_hook('pre_pull', extras, GitMessageWriter(stdout))
 
     except Exception as error:
         log.exception('Failed to call pre_pull hook')
         status_code = 128
         stdout.write(f'ERROR: {error}\n')
 
     return HookResponse(status_code, stdout.getvalue())
 
 
 def git_post_pull(extras) -> HookResponse:
     """
     Post pull hook.
 
     :param extras: dictionary containing the keys defined in simplevcs
     :type extras: dict
 
     :return: status code of the hook. 0 for success.
     :rtype: int
     """
     if 'pull' not in extras['hooks']:
         return HookResponse(0, '')
 
     stdout = io.StringIO()
     try:
         status = _call_hook('post_pull', extras, GitMessageWriter(stdout))
     except Exception as error:
         status = 128
         stdout.write(f'ERROR: {error}\n')
 
     return HookResponse(status, stdout.getvalue())
 
 
 def _parse_git_ref_lines(revision_lines):
     rev_data = []
     for revision_line in revision_lines or []:
         old_rev, new_rev, ref = revision_line.strip().split(' ')
         ref_data = ref.split('/', 2)
         if ref_data[1] in ('tags', 'heads'):
             rev_data.append({
                 # NOTE(marcink):
                 # we're unable to tell total_commits for git at this point
                 # but we set the variable for consistency with GIT
                 'total_commits': -1,
                 'old_rev': old_rev,
                 'new_rev': new_rev,
                 'ref': ref,
                 'type': ref_data[1],
                 'name': ref_data[2],
             })
     return rev_data
 
 
 def git_pre_receive(unused_repo_path, revision_lines, env) -> int:
     """
     Pre push hook.
 
     :return: status code of the hook. 0 for success.
     """
     extras = json.loads(env['RC_SCM_DATA'])
     rev_data = _parse_git_ref_lines(revision_lines)
     if 'push' not in extras['hooks']:
         return 0
     _fix_hooks_executables(env.get('RC_INI_FILE'))
 
     empty_commit_id = '0' * 40
 
     detect_force_push = extras.get('detect_force_push')
 
     for push_ref in rev_data:
         # store our git-env which holds the temp store
         push_ref['git_env'] = _get_git_env()
         push_ref['pruned_sha'] = ''
         if not detect_force_push:
             # don't check for forced-push when we don't need to
             continue
 
         type_ = push_ref['type']
         new_branch = push_ref['old_rev'] == empty_commit_id
         delete_branch = push_ref['new_rev'] == empty_commit_id
         if type_ == 'heads' and not (new_branch or delete_branch):
             old_rev = push_ref['old_rev']
             new_rev = push_ref['new_rev']
             cmd = [settings.GIT_EXECUTABLE(), 'rev-list', old_rev, f'^{new_rev}']
             stdout, stderr = subprocessio.run_command(
                 cmd, env=os.environ.copy())
             # means we're having some non-reachable objects, this forced push was used
             if stdout:
                 push_ref['pruned_sha'] = stdout.splitlines()
 
     extras['hook_type'] = 'pre_receive'
     extras['commit_ids'] = rev_data
 
     stdout = sys.stdout
     status_code = _call_hook('pre_push', extras, GitMessageWriter(stdout))
 
     return status_code
 
 
 def git_post_receive(unused_repo_path, revision_lines, env) -> int:
     """
     Post push hook.
 
     :return: status code of the hook. 0 for success.
     """
     extras = json.loads(env['RC_SCM_DATA'])
     if 'push' not in extras['hooks']:
         return 0
 
     _fix_hooks_executables(env.get('RC_INI_FILE'))
 
     rev_data = _parse_git_ref_lines(revision_lines)
 
     git_revs = []
 
     # N.B.(skreft): it is ok to just call git, as git before calling a
     # subcommand sets the PATH environment variable so that it point to the
     # correct version of the git executable.
     empty_commit_id = '0' * 40
     branches = []
     tags = []
     for push_ref in rev_data:
         type_ = push_ref['type']
 
         if type_ == 'heads':
             # starting new branch case
             if push_ref['old_rev'] == empty_commit_id:
                 push_ref_name = push_ref['name']
 
                 if push_ref_name not in branches:
                     branches.append(push_ref_name)
 
                 need_head_set = ''
                 with Repository(os.getcwd()) as repo:
                     try:
                         repo.head
                     except pygit2.GitError:
                         need_head_set = f'refs/heads/{push_ref_name}'
 
                     if need_head_set:
                         repo.set_head(need_head_set)
                         print(f"Setting default branch to {push_ref_name}")
 
                 cmd = [settings.GIT_EXECUTABLE(), 'for-each-ref', '--format=%(refname)', 'refs/heads/*']
                 stdout, stderr = subprocessio.run_command(
                     cmd, env=os.environ.copy())
                 heads = safe_str(stdout)
                 heads = heads.replace(push_ref['ref'], '')
                 heads = ' '.join(head for head
                                  in heads.splitlines() if head) or '.'
                 cmd = [settings.GIT_EXECUTABLE(), 'log', '--reverse',
                        '--pretty=format:%H', '--', push_ref['new_rev'],
                        '--not', heads]
                 stdout, stderr = subprocessio.run_command(
                     cmd, env=os.environ.copy())
                 git_revs.extend(list(map(ascii_str, stdout.splitlines())))
 
             # delete branch case
             elif push_ref['new_rev'] == empty_commit_id:
                 git_revs.append(f'delete_branch=>{push_ref["name"]}')
             else:
                 if push_ref['name'] not in branches:
                     branches.append(push_ref['name'])
 
                 cmd = [settings.GIT_EXECUTABLE(), 'log',
                        f'{push_ref["old_rev"]}..{push_ref["new_rev"]}',
                        '--reverse', '--pretty=format:%H']
                 stdout, stderr = subprocessio.run_command(
                     cmd, env=os.environ.copy())
                 # we get bytes from stdout, we need str to be consistent
                 log_revs = list(map(ascii_str, stdout.splitlines()))
                 git_revs.extend(log_revs)
 
                 # Pure pygit2 impl. but still 2-3x slower :/
                 # results = []
                 #
                 # with Repository(os.getcwd()) as repo:
                 #     repo_new_rev = repo[push_ref['new_rev']]
                 #     repo_old_rev = repo[push_ref['old_rev']]
                 #     walker = repo.walk(repo_new_rev.id, pygit2.GIT_SORT_TOPOLOGICAL)
                 #
                 #     for commit in walker:
                 #         if commit.id == repo_old_rev.id:
                 #             break
                 #         results.append(commit.id.hex)
                 #     # reverse the order, can't use GIT_SORT_REVERSE
                 #     log_revs = results[::-1]
 
         elif type_ == 'tags':
             if push_ref['name'] not in tags:
                 tags.append(push_ref['name'])
             git_revs.append(f'tag=>{push_ref["name"]}')
 
     extras['hook_type'] = 'post_receive'
     extras['commit_ids'] = git_revs
     extras['new_refs'] = {
         'branches': branches,
         'bookmarks': [],
         'tags': tags,
     }
 
     stdout = sys.stdout
 
     if 'repo_size' in extras['hooks']:
         try:
             _call_hook('repo_size', extras, GitMessageWriter(stdout))
         except Exception:
             pass
 
     status_code = _call_hook('post_push', extras, GitMessageWriter(stdout))
     return status_code
 
 
 def get_extras_from_txn_id(repo_path, txn_id):
     extras = get_txn_id_from_store(repo_path, txn_id)
     return extras
 
 
 def svn_pre_commit(repo_path, commit_data, env):
 
     path, txn_id = commit_data
     branches = []
     tags = []
 
     if env.get('RC_SCM_DATA'):
         extras = json.loads(env['RC_SCM_DATA'])
     else:
         ini_path = env.get('RC_INI_FILE')
         if ini_path:
             _get_ini_settings(ini_path)
         # fallback method to read from TXN-ID stored data
         extras = get_extras_from_txn_id(path, txn_id)
 
     if not extras:
         raise ValueError('SVN-PRE-COMMIT: Failed to extract context data in called extras for hook execution')
 
     if extras.get('rc_internal_commit'):
         # special marker for internal commit, we don't call hooks client
         return 0
 
     extras['hook_type'] = 'pre_commit'
     extras['commit_ids'] = [txn_id]
     extras['txn_id'] = txn_id
     extras['new_refs'] = {
         'total_commits': 1,
         'branches': branches,
         'bookmarks': [],
         'tags': tags,
     }
 
     return _call_hook('pre_push', extras, SvnMessageWriter())
 
 
 def svn_post_commit(repo_path, commit_data, env):
     """
     commit_data is path, rev, txn_id
     """
 
     if len(commit_data) == 3:
         path, commit_id, txn_id = commit_data
     elif len(commit_data) == 2:
         log.error('Failed to extract txn_id from commit_data using legacy method. '
                   'Some functionality might be limited')
         path, commit_id = commit_data
         txn_id = None
     else:
         return 0
 
     branches = []
     tags = []
 
     if env.get('RC_SCM_DATA'):
         extras = json.loads(env['RC_SCM_DATA'])
     else:
         ini_path = env.get('RC_INI_FILE')
         if ini_path:
             _get_ini_settings(ini_path)
         # fallback method to read from TXN-ID stored data
         extras = get_extras_from_txn_id(path, txn_id)
 
     if not extras and txn_id:
         raise ValueError('SVN-POST-COMMIT: Failed to extract context data in called extras for hook execution')
 
     if extras.get('rc_internal_commit'):
         # special marker for internal commit, we don't call hooks client
         return 0
 
     extras['hook_type'] = 'post_commit'
     extras['commit_ids'] = [commit_id]
     extras['txn_id'] = txn_id
     extras['new_refs'] = {
         'branches': branches,
         'bookmarks': [],
         'tags': tags,
         'total_commits': 1,
     }
 
     if 'repo_size' in extras['hooks']:
         try:
             _call_hook('repo_size', extras, SvnMessageWriter())
         except Exception:
             pass
 
     return _call_hook('post_push', extras, SvnMessageWriter())

vcsserver/http_main.py

@@ -1,763 +1,765 b''
 # RhodeCode VCSServer provides access to different vcs backends via network.
 # Copyright (C) 2014-2023 RhodeCode GmbH
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
 # the Free Software Foundation; either version 3 of the License, or
 # (at your option) any later version.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
 # along with this program; if not, write to the Free Software Foundation,
 # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301  USA
 
 import io
 import os
 import platform
 import sys
 import locale
 import logging
 import uuid
 import time
 import wsgiref.util
 import tempfile
 import psutil
 
 from itertools import chain
 
 import msgpack
 import configparser
 
 from pyramid.config import Configurator
 from pyramid.wsgi import wsgiapp
 from pyramid.response import Response
 
 from vcsserver.base import BytesEnvelope, BinaryEnvelope
 
 from vcsserver.config.settings_maker import SettingsMaker
 
 from vcsserver.tweens.request_wrapper import get_headers_call_context
 
 from vcsserver import remote_wsgi, scm_app, hgpatches
 from vcsserver.server import VcsServer
 from vcsserver.git_lfs.app import GIT_LFS_CONTENT_TYPE, GIT_LFS_PROTO_PAT
 from vcsserver.echo_stub import remote_wsgi as remote_wsgi_stub
 from vcsserver.echo_stub.echo_app import EchoApp
-from vcsserver.exceptions import HTTPRepoLocked, HTTPRepoBranchProtected
+from vcsserver.exceptions import HTTPRepoLocked, HTTPRepoBranchProtected, HTTPClientNotSupported
 from vcsserver.lib.exc_tracking import store_exception, format_exc
 from vcsserver.lib.str_utils import safe_int
 from vcsserver.lib.statsd_client import StatsdClient
 from vcsserver.lib.ext_json import json
 from vcsserver.lib.config_utils import configure_and_store_settings
 
 
 strict_vcs = True
 
 git_import_err = None
 try:
     from vcsserver.remote.git_remote import GitFactory, GitRemote
 except ImportError as e:
     GitFactory = None
     GitRemote = None
     git_import_err = e
     if strict_vcs:
         raise
 
 
 hg_import_err = None
 try:
     from vcsserver.remote.hg_remote import MercurialFactory, HgRemote
 except ImportError as e:
     MercurialFactory = None
     HgRemote = None
     hg_import_err = e
     if strict_vcs:
         raise
 
 
 svn_import_err = None
 try:
     from vcsserver.remote.svn_remote import SubversionFactory, SvnRemote
 except ImportError as e:
     SubversionFactory = None
     SvnRemote = None
     svn_import_err = e
     if strict_vcs:
         raise
 
 log = logging.getLogger(__name__)
 
 # due to Mercurial/glibc2.27 problems we need to detect if locale settings are
 # causing problems and "fix" it in case they do and fallback to LC_ALL = C
 
 try:
     locale.setlocale(locale.LC_ALL, '')
 except locale.Error as e:
     log.error('LOCALE ERROR: failed to set LC_ALL, fallback to LC_ALL=C, org error: %s', e)
     os.environ['LC_ALL'] = 'C'
 
 
 def _is_request_chunked(environ):
     stream = environ.get('HTTP_TRANSFER_ENCODING', '') == 'chunked'
     return stream
 
 
 def log_max_fd():
     try:
         maxfd = psutil.Process().rlimit(psutil.RLIMIT_NOFILE)[1]
         log.info('Max file descriptors value: %s', maxfd)
     except Exception:
         pass
 
 
 class VCS:
     def __init__(self, locale_conf=None, cache_config=None):
         self.locale = locale_conf
         self.cache_config = cache_config
         self._configure_locale()
 
         log_max_fd()
 
         if GitFactory and GitRemote:
             git_factory = GitFactory()
             self._git_remote = GitRemote(git_factory)
         else:
             log.error("Git client import failed: %s", git_import_err)
 
         if MercurialFactory and HgRemote:
             hg_factory = MercurialFactory()
             self._hg_remote = HgRemote(hg_factory)
         else:
             log.error("Mercurial client import failed: %s", hg_import_err)
 
         if SubversionFactory and SvnRemote:
             svn_factory = SubversionFactory()
 
             # hg factory is used for svn url validation
             hg_factory = MercurialFactory()
             self._svn_remote = SvnRemote(svn_factory, hg_factory=hg_factory)
         else:
             log.error("Subversion client import failed: %s", svn_import_err)
 
         self._vcsserver = VcsServer()
 
     def _configure_locale(self):
         if self.locale:
             log.info('Settings locale: `LC_ALL` to %s', self.locale)
         else:
             log.info('Configuring locale subsystem based on environment variables')
         try:
             # If self.locale is the empty string, then the locale
             # module will use the environment variables. See the
             # documentation of the package `locale`.
             locale.setlocale(locale.LC_ALL, self.locale)
 
             language_code, encoding = locale.getlocale()
             log.info(
                 'Locale set to language code "%s" with encoding "%s".',
                 language_code, encoding)
         except locale.Error:
             log.exception('Cannot set locale, not configuring the locale system')
 
 
 class WsgiProxy:
     def __init__(self, wsgi):
         self.wsgi = wsgi
 
     def __call__(self, environ, start_response):
         input_data = environ['wsgi.input'].read()
         input_data = msgpack.unpackb(input_data)
 
         error = None
         try:
             data, status, headers = self.wsgi.handle(
                 input_data['environment'], input_data['input_data'],
                 *input_data['args'], **input_data['kwargs'])
         except Exception as e:
             data, status, headers = [], None, None
             error = {
                 'message': str(e),
                 '_vcs_kind': getattr(e, '_vcs_kind', None)
             }
 
         start_response(200, {})
         return self._iterator(error, status, headers, data)
 
     def _iterator(self, error, status, headers, data):
         initial_data = [
             error,
             status,
             headers,
         ]
 
         for d in chain(initial_data, data):
             yield msgpack.packb(d)
 
 
 def not_found(request):
     return {'status': '404 NOT FOUND'}
 
 
 class VCSViewPredicate:
     def __init__(self, val, config):
         self.remotes = val
 
     def text(self):
         return f'vcs view method = {list(self.remotes.keys())}'
 
     phash = text
 
     def __call__(self, context, request):
         """
         View predicate that returns true if given backend is supported by
         defined remotes.
         """
         backend = request.matchdict.get('backend')
         return backend in self.remotes
 
 
 class HTTPApplication:
     ALLOWED_EXCEPTIONS = ('KeyError', 'URLError')
 
     remote_wsgi = remote_wsgi
     _use_echo_app = False
 
     def __init__(self, settings=None, global_config=None):
 
         self.config = Configurator(settings=settings)
         # Init our statsd at very start
         self.config.registry.statsd = StatsdClient.statsd
         self.config.registry.vcs_call_context = {}
 
         self.global_config = global_config
         self.config.include('vcsserver.lib.rc_cache')
         self.config.include('vcsserver.lib.archive_cache')
 
         settings_locale = settings.get('locale', '') or 'en_US.UTF-8'
         vcs = VCS(locale_conf=settings_locale, cache_config=settings)
         self._remotes = {
             'hg': vcs._hg_remote,
             'git': vcs._git_remote,
             'svn': vcs._svn_remote,
             'server': vcs._vcsserver,
         }
         if settings.get('dev.use_echo_app', 'false').lower() == 'true':
             self._use_echo_app = True
             log.warning("Using EchoApp for VCS operations.")
             self.remote_wsgi = remote_wsgi_stub
 
         configure_and_store_settings(global_config, settings)
 
         self._configure()
 
     def _configure(self):
         self.config.add_renderer(name='msgpack', factory=self._msgpack_renderer_factory)
 
         self.config.add_route('service', '/_service')
         self.config.add_route('status', '/status')
         self.config.add_route('hg_proxy', '/proxy/hg')
         self.config.add_route('git_proxy', '/proxy/git')
 
         # rpc methods
         self.config.add_route('vcs', '/{backend}')
 
         # streaming rpc remote methods
         self.config.add_route('vcs_stream', '/{backend}/stream')
 
         # vcs operations clone/push as streaming
         self.config.add_route('stream_git', '/stream/git/*repo_name')
         self.config.add_route('stream_hg', '/stream/hg/*repo_name')
 
         self.config.add_view(self.status_view, route_name='status', renderer='json')
         self.config.add_view(self.service_view, route_name='service', renderer='msgpack')
 
         self.config.add_view(self.hg_proxy(), route_name='hg_proxy')
         self.config.add_view(self.git_proxy(), route_name='git_proxy')
         self.config.add_view(self.vcs_view, route_name='vcs', renderer='msgpack',
                              vcs_view=self._remotes)
         self.config.add_view(self.vcs_stream_view, route_name='vcs_stream',
                              vcs_view=self._remotes)
 
         self.config.add_view(self.hg_stream(), route_name='stream_hg')
         self.config.add_view(self.git_stream(), route_name='stream_git')
 
         self.config.add_view_predicate('vcs_view', VCSViewPredicate)
 
         self.config.add_notfound_view(not_found, renderer='json')
 
         self.config.add_view(self.handle_vcs_exception, context=Exception)
 
         self.config.add_tween(
             'vcsserver.tweens.request_wrapper.RequestWrapperTween',
         )
         self.config.add_request_method(
             'vcsserver.lib.request_counter.get_request_counter',
             'request_count')
 
     def wsgi_app(self):
         return self.config.make_wsgi_app()
 
     def _vcs_view_params(self, request):
         remote = self._remotes[request.matchdict['backend']]
         payload = msgpack.unpackb(request.body, use_list=True)
 
         method = payload.get('method')
         params = payload['params']
         wire = params.get('wire')
         args = params.get('args')
         kwargs = params.get('kwargs')
         context_uid = None
 
         request.registry.vcs_call_context = {
             'method': method,
             'repo_name': payload.get('_repo_name'),
         }
 
         if wire:
             try:
                 wire['context'] = context_uid = uuid.UUID(wire['context'])
             except KeyError:
                 pass
             args.insert(0, wire)
         repo_state_uid = wire.get('repo_state_uid') if wire else None
 
         # NOTE(marcink): trading complexity for slight performance
         if log.isEnabledFor(logging.DEBUG):
             # also we SKIP printing out any of those methods args since they maybe excessive
             just_args_methods = {
                 'commitctx': ('content', 'removed', 'updated'),
                 'commit': ('content', 'removed', 'updated')
             }
             if method in just_args_methods:
                 skip_args = just_args_methods[method]
                 call_args = ''
                 call_kwargs = {}
                 for k in kwargs:
                     if k in skip_args:
                         # replace our skip key with dummy
                         call_kwargs[k] = f'RemovedParam({k})'
                     else:
                         call_kwargs[k] = kwargs[k]
             else:
                 call_args = args[1:]
                 call_kwargs = kwargs
 
             log.debug('Method requested:`%s` with args:%s kwargs:%s context_uid: %s, repo_state_uid:%s',
                       method, call_args, call_kwargs, context_uid, repo_state_uid)
 
         statsd = request.registry.statsd
         if statsd:
             statsd.incr(
                 'vcsserver_method_total', tags=[
                     f"method:{method}",
                 ])
         return payload, remote, method, args, kwargs
 
     def vcs_view(self, request):
 
         payload, remote, method, args, kwargs = self._vcs_view_params(request)
         payload_id = payload.get('id')
 
         try:
             resp = getattr(remote, method)(*args, **kwargs)
         except Exception as e:
             exc_info = list(sys.exc_info())
             exc_type, exc_value, exc_traceback = exc_info
 
             org_exc = getattr(e, '_org_exc', None)
             org_exc_name = None
             org_exc_tb = ''
             if org_exc:
                 org_exc_name = org_exc.__class__.__name__
                 org_exc_tb = getattr(e, '_org_exc_tb', '')
                 # replace our "faked" exception with our org
                 exc_info[0] = org_exc.__class__
                 exc_info[1] = org_exc
 
             should_store_exc = True
             if org_exc:
                 def get_exc_fqn(_exc_obj):
                     module_name = getattr(org_exc.__class__, '__module__', 'UNKNOWN')
                     return module_name + '.' + org_exc_name
 
                 exc_fqn = get_exc_fqn(org_exc)
 
                 if exc_fqn in ['mercurial.error.RepoLookupError',
                                'vcsserver.exceptions.RefNotFoundException']:
                     should_store_exc = False
 
             if should_store_exc:
                 store_exception(id(exc_info), exc_info, request_path=request.path)
 
             tb_info = format_exc(exc_info)
 
             type_ = e.__class__.__name__
             if type_ not in self.ALLOWED_EXCEPTIONS:
                 type_ = None
 
             resp = {
                 'id': payload_id,
                 'error': {
                     'message': str(e),
                     'traceback': tb_info,
                     'org_exc': org_exc_name,
                     'org_exc_tb': org_exc_tb,
                     'type': type_
                 }
             }
 
             try:
                 resp['error']['_vcs_kind'] = getattr(e, '_vcs_kind', None)
             except AttributeError:
                 pass
         else:
             resp = {
                 'id': payload_id,
                 'result': resp
             }
         log.debug('Serving data for method %s', method)
         return resp
 
     def vcs_stream_view(self, request):
         payload, remote, method, args, kwargs = self._vcs_view_params(request)
         # this method has a stream: marker we remove it here
         method = method.split('stream:')[-1]
         chunk_size = safe_int(payload.get('chunk_size')) or 4096
 
         resp = getattr(remote, method)(*args, **kwargs)
 
         def get_chunked_data(method_resp):
             stream = io.BytesIO(method_resp)
             while 1:
                 chunk = stream.read(chunk_size)
                 if not chunk:
                     break
                 yield chunk
 
         response = Response(app_iter=get_chunked_data(resp))
         response.content_type = 'application/octet-stream'
 
         return response
 
     def status_view(self, request):
         import vcsserver
         _platform_id = platform.uname()[1] or 'instance'
 
         return {
             "status": "OK",
             "vcsserver_version": vcsserver.get_version(),
             "platform": _platform_id,
             "pid": os.getpid(),
         }
 
     def service_view(self, request):
         import vcsserver
 
         payload = msgpack.unpackb(request.body, use_list=True)
         server_config, app_config = {}, {}
 
         try:
             path = self.global_config['__file__']
             config = configparser.RawConfigParser()
 
             config.read(path)
 
             if config.has_section('server:main'):
                 server_config = dict(config.items('server:main'))
             if config.has_section('app:main'):
                 app_config = dict(config.items('app:main'))
 
         except Exception:
             log.exception('Failed to read .ini file for display')
 
         environ = list(os.environ.items())
 
         resp = {
             'id': payload.get('id'),
             'result': dict(
                 version=vcsserver.get_version(),
                 config=server_config,
                 app_config=app_config,
                 environ=environ,
                 payload=payload,
             )
         }
         return resp
 
     def _msgpack_renderer_factory(self, info):
 
         def _render(value, system):
             bin_type = False
             res = value.get('result')
             if isinstance(res, BytesEnvelope):
                 log.debug('Result is wrapped in BytesEnvelope type')
                 bin_type = True
             elif isinstance(res, BinaryEnvelope):
                 log.debug('Result is wrapped in BinaryEnvelope type')
                 value['result'] = res.val
                 bin_type = True
 
             request = system.get('request')
             if request is not None:
                 response = request.response
                 ct = response.content_type
                 if ct == response.default_content_type:
                     response.content_type = 'application/x-msgpack'
                     if bin_type:
                         response.content_type = 'application/x-msgpack-bin'
 
             return msgpack.packb(value, use_bin_type=bin_type)
         return _render
 
     def set_env_from_config(self, environ, config):
         dict_conf = {}
         try:
             for elem in config:
                 if elem[0] == 'rhodecode':
                     dict_conf = json.loads(elem[2])
                     break
         except Exception:
             log.exception('Failed to fetch SCM CONFIG')
             return
 
         username = dict_conf.get('username')
         if username:
             environ['REMOTE_USER'] = username
             # mercurial specific, some extension api rely on this
             environ['HGUSER'] = username
 
         ip = dict_conf.get('ip')
         if ip:
             environ['REMOTE_HOST'] = ip
 
         if _is_request_chunked(environ):
             # set the compatibility flag for webob
             environ['wsgi.input_terminated'] = True
 
     def hg_proxy(self):
         @wsgiapp
         def _hg_proxy(environ, start_response):
             app = WsgiProxy(self.remote_wsgi.HgRemoteWsgi())
             return app(environ, start_response)
         return _hg_proxy
 
     def git_proxy(self):
         @wsgiapp
         def _git_proxy(environ, start_response):
             app = WsgiProxy(self.remote_wsgi.GitRemoteWsgi())
             return app(environ, start_response)
         return _git_proxy
 
     def hg_stream(self):
         if self._use_echo_app:
             @wsgiapp
             def _hg_stream(environ, start_response):
                 app = EchoApp('fake_path', 'fake_name', None)
                 return app(environ, start_response)
             return _hg_stream
         else:
             @wsgiapp
             def _hg_stream(environ, start_response):
                 log.debug('http-app: handling hg stream')
                 call_context = get_headers_call_context(environ)
 
                 repo_path = call_context['repo_path']
                 repo_name = call_context['repo_name']
                 config = call_context['repo_config']
 
                 app = scm_app.create_hg_wsgi_app(
                     repo_path, repo_name, config)
 
                 # Consistent path information for hgweb
                 environ['PATH_INFO'] = call_context['path_info']
                 environ['REPO_NAME'] = repo_name
                 self.set_env_from_config(environ, config)
 
                 log.debug('http-app: starting app handler '
                           'with %s and process request', app)
                 return app(environ, ResponseFilter(start_response))
             return _hg_stream
 
     def git_stream(self):
         if self._use_echo_app:
             @wsgiapp
             def _git_stream(environ, start_response):
                 app = EchoApp('fake_path', 'fake_name', None)
                 return app(environ, start_response)
             return _git_stream
         else:
             @wsgiapp
             def _git_stream(environ, start_response):
                 log.debug('http-app: handling git stream')
 
                 call_context = get_headers_call_context(environ)
 
                 repo_path = call_context['repo_path']
                 repo_name = call_context['repo_name']
                 config = call_context['repo_config']
 
                 environ['PATH_INFO'] = call_context['path_info']
                 self.set_env_from_config(environ, config)
 
                 content_type = environ.get('CONTENT_TYPE', '')
 
                 path = environ['PATH_INFO']
                 is_lfs_request = GIT_LFS_CONTENT_TYPE in content_type
                 log.debug(
                     'LFS: Detecting if request `%s` is LFS server path based '
                     'on content type:`%s`, is_lfs:%s',
                     path, content_type, is_lfs_request)
 
                 if not is_lfs_request:
                     # fallback detection by path
                     if GIT_LFS_PROTO_PAT.match(path):
                         is_lfs_request = True
                     log.debug(
                         'LFS: fallback detection by path of: `%s`, is_lfs:%s',
                         path, is_lfs_request)
 
                 if is_lfs_request:
                     app = scm_app.create_git_lfs_wsgi_app(
                         repo_path, repo_name, config)
                 else:
                     app = scm_app.create_git_wsgi_app(
                         repo_path, repo_name, config)
 
                 log.debug('http-app: starting app handler '
                           'with %s and process request', app)
 
                 return app(environ, start_response)
 
             return _git_stream
 
     def handle_vcs_exception(self, exception, request):
-        _vcs_kind = getattr(exception, '_vcs_kind', '')
 
-        if _vcs_kind == 'repo_locked':
-            headers_call_context = get_headers_call_context(request.environ)
-            status_code = safe_int(headers_call_context['locked_status_code'])
+        match _vcs_kind := getattr(exception, '_vcs_kind', ''):
+            case 'repo_locked':
+                headers_call_context = get_headers_call_context(request.environ)
+                status_code = safe_int(headers_call_context['locked_status_code'])
 
-            return HTTPRepoLocked(
-                title=str(exception), status_code=status_code, headers=[('X-Rc-Locked', '1')])
-
-        elif _vcs_kind == 'repo_branch_protected':
-            # Get custom repo-branch-protected status code if present.
-            return HTTPRepoBranchProtected(
-                title=str(exception), headers=[('X-Rc-Branch-Protection', '1')])
+                return HTTPRepoLocked(
+                    title=str(exception), status_code=status_code, headers=[('X-Rc-Locked', '1')])
+            case 'repo_branch_protected':
+                # Get custom repo-branch-protected status code if present.
+                return HTTPRepoBranchProtected(
+                    title=str(exception), headers=[('X-Rc-Branch-Protection', '1')])
+            case 'client_not_supported':
+                return HTTPClientNotSupported(
+                    title=str(exception), headers=[('X-Rc-Client-Not-Supported', '1')])
 
         exc_info = request.exc_info
         store_exception(id(exc_info), exc_info)
 
         traceback_info = 'unavailable'
         if request.exc_info:
             traceback_info = format_exc(request.exc_info)
 
         log.error(
             'error occurred handling this request for path: %s, \n%s',
             request.path, traceback_info)
 
         statsd = request.registry.statsd
         if statsd:
             exc_type = f"{exception.__class__.__module__}.{exception.__class__.__name__}"
             statsd.incr('vcsserver_exception_total',
                         tags=[f"type:{exc_type}"])
         raise exception
 
 
 class ResponseFilter:
 
     def __init__(self, start_response):
         self._start_response = start_response
 
     def __call__(self, status, response_headers, exc_info=None):
         headers = tuple(
             (h, v) for h, v in response_headers
             if not wsgiref.util.is_hop_by_hop(h))
         return self._start_response(status, headers, exc_info)
 
 
 def sanitize_settings_and_apply_defaults(global_config, settings):
     _global_settings_maker = SettingsMaker(global_config)
     settings_maker = SettingsMaker(settings)
 
     settings_maker.make_setting('logging.autoconfigure', False, parser='bool')
 
     logging_conf = os.path.join(os.path.dirname(global_config.get('__file__')), 'logging.ini')
     settings_maker.enable_logging(logging_conf)
 
     # Default includes, possible to change as a user
     pyramid_includes = settings_maker.make_setting('pyramid.includes', [], parser='list:newline')
     log.debug("Using the following pyramid.includes: %s", pyramid_includes)
 
     settings_maker.make_setting('__file__', global_config.get('__file__'))
 
     settings_maker.make_setting('pyramid.default_locale_name', 'en')
     settings_maker.make_setting('locale', 'en_US.UTF-8')
 
     settings_maker.make_setting(
         'core.binary_dir', '/usr/local/bin/rhodecode_bin/vcs_bin',
         default_when_empty=True, parser='string:noquote')
 
     settings_maker.make_setting('vcs.svn.redis_conn', 'redis://redis:6379/0')
 
     temp_store = tempfile.gettempdir()
     default_cache_dir = os.path.join(temp_store, 'rc_cache')
     # save default, cache dir, and use it for all backends later.
     default_cache_dir = settings_maker.make_setting(
         'cache_dir',
         default=default_cache_dir, default_when_empty=True,
         parser='dir:ensured')
 
     # exception store cache
     settings_maker.make_setting(
         'exception_tracker.store_path',
         default=os.path.join(default_cache_dir, 'exc_store'), default_when_empty=True,
         parser='dir:ensured'
     )
 
     # repo_object cache defaults
     settings_maker.make_setting(
         'rc_cache.repo_object.backend',
         default='dogpile.cache.rc.file_namespace',
         parser='string')
     settings_maker.make_setting(
         'rc_cache.repo_object.expiration_time',
         default=30 * 24 * 60 * 60,  # 30days
         parser='int')
     settings_maker.make_setting(
         'rc_cache.repo_object.arguments.filename',
         default=os.path.join(default_cache_dir, 'vcsserver_cache_repo_object.db'),
         parser='string')
 
     # statsd
     settings_maker.make_setting('statsd.enabled', False, parser='bool')
     settings_maker.make_setting('statsd.statsd_host', 'statsd-exporter', parser='string')
     settings_maker.make_setting('statsd.statsd_port', 9125, parser='int')
     settings_maker.make_setting('statsd.statsd_prefix', '')
     settings_maker.make_setting('statsd.statsd_ipv6', False, parser='bool')
 
     settings_maker.env_expand()
 
 
 def main(global_config, **settings):
     start_time = time.time()
     log.info('Pyramid app config starting')
 
     if MercurialFactory:
         hgpatches.patch_largefiles_capabilities()
         hgpatches.patch_subrepo_type_mapping()
 
     # Fill in and sanitize the defaults & do ENV expansion
     sanitize_settings_and_apply_defaults(global_config, settings)
 
     # init and bootstrap StatsdClient
     StatsdClient.setup(settings)
 
     pyramid_app = HTTPApplication(settings=settings, global_config=global_config).wsgi_app()
     total_time = time.time() - start_time
     log.info('Pyramid app created and configured in %.2fs', total_time)
     return pyramid_app

vcsserver/subprocessio.py

@@ -1,563 +1,564 b''
 """
 Module provides a class allowing to wrap communication over subprocess.Popen
 input, output, error streams into a meaningfull, non-blocking, concurrent
 stream processor exposing the output data as an iterator fitting to be a
 return value passed by a WSGI applicaiton to a WSGI server per PEP 3333.
 
 Copyright (c) 2011  Daniel Dotsenko <dotsa[at]hotmail.com>
 
 This file is part of git_http_backend.py Project.
 
 git_http_backend.py Project is free software: you can redistribute it and/or
 modify it under the terms of the GNU Lesser General Public License as
 published by the Free Software Foundation, either version 2.1 of the License,
 or (at your option) any later version.
 
 git_http_backend.py Project is distributed in the hope that it will be useful,
 but WITHOUT ANY WARRANTY; without even the implied warranty of
 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
 GNU Lesser General Public License for more details.
 
 You should have received a copy of the GNU Lesser General Public License
 along with git_http_backend.py Project.
 If not, see <http://www.gnu.org/licenses/>.
 """
 import os
 import collections
 import logging
 import subprocess
 import threading
 
 from vcsserver.lib.str_utils import safe_str
 
 log = logging.getLogger(__name__)
 
 
 class StreamFeeder(threading.Thread):
     """
     Normal writing into pipe-like is blocking once the buffer is filled.
     This thread allows a thread to seep data from a file-like into a pipe
     without blocking the main thread.
     We close inpipe once the end of the source stream is reached.
     """
 
     def __init__(self, source):
         super().__init__()
         self.daemon = True
         filelike = False
         self.bytes = b''
         if type(source) in (str, bytes, bytearray):  # string-like
             self.bytes = bytes(source)
         else:  # can be either file pointer or file-like
             if isinstance(source, int):  # file pointer it is
                 # converting file descriptor (int) stdin into file-like
                 source = os.fdopen(source, 'rb', 16384)
             # let's see if source is file-like by now
             filelike = hasattr(source, 'read')
         if not filelike and not self.bytes:
             raise TypeError("StreamFeeder's source object must be a readable "
                             "file-like, a file descriptor, or a string-like.")
         self.source = source
         self.readiface, self.writeiface = os.pipe()
 
     def run(self):
         writer = self.writeiface
         try:
             if self.bytes:
                 os.write(writer, self.bytes)
             else:
                 s = self.source
 
                 while 1:
                     _bytes = s.read(4096)
                     if not _bytes:
                         break
                     os.write(writer, _bytes)
 
         finally:
             os.close(writer)
 
     @property
     def output(self):
         return self.readiface
 
 
 class InputStreamChunker(threading.Thread):
     def __init__(self, source, target, buffer_size, chunk_size):
 
         super().__init__()
 
         self.daemon = True  # die die die.
 
         self.source = source
         self.target = target
         self.chunk_count_max = int(buffer_size / chunk_size) + 1
         self.chunk_size = chunk_size
 
         self.data_added = threading.Event()
         self.data_added.clear()
 
         self.keep_reading = threading.Event()
         self.keep_reading.set()
 
         self.EOF = threading.Event()
         self.EOF.clear()
 
         self.go = threading.Event()
         self.go.set()
 
     def stop(self):
         self.go.clear()
         self.EOF.set()
         try:
             # this is not proper, but is done to force the reader thread let
             # go of the input because, if successful, .close() will send EOF
             # down the pipe.
             self.source.close()
         except Exception:
             pass
 
     def run(self):
         s = self.source
         t = self.target
         cs = self.chunk_size
         chunk_count_max = self.chunk_count_max
         keep_reading = self.keep_reading
         da = self.data_added
         go = self.go
 
         try:
             b = s.read(cs)
         except ValueError:
             b = ''
 
         timeout_input = 20
         while b and go.is_set():
             if len(t) > chunk_count_max:
                 keep_reading.clear()
                 keep_reading.wait(timeout_input)
                 if len(t) > chunk_count_max + timeout_input:
                     log.error("Timed out while waiting for input from subprocess.")
                     os._exit(-1)  # this will cause the worker to recycle itself
 
             t.append(b)
             da.set()
 
             try:
                 b = s.read(cs)
             except ValueError:  # probably "I/O operation on closed file"
                 b = ''
 
         self.EOF.set()
         da.set()  # for cases when done but there was no input.
 
 
 class BufferedGenerator:
     """
     Class behaves as a non-blocking, buffered pipe reader.
     Reads chunks of data (through a thread)
     from a blocking pipe, and attaches these to an array (Deque) of chunks.
     Reading is halted in the thread when max chunks is internally buffered.
     The .next() may operate in blocking or non-blocking fashion by yielding
     '' if no data is ready
     to be sent or by not returning until there is some data to send
     When we get EOF from underlying source pipe we raise the marker to raise
     StopIteration after the last chunk of data is yielded.
     """
 
     def __init__(self, name, source, buffer_size=65536, chunk_size=4096,
                  starting_values=None, bottomless=False):
         starting_values = starting_values or []
         self.name = name
         self.buffer_size = buffer_size
         self.chunk_size = chunk_size
 
         if bottomless:
             maxlen = int(buffer_size / chunk_size)
         else:
             maxlen = None
 
         self.data_queue = collections.deque(starting_values, maxlen)
         self.worker = InputStreamChunker(source, self.data_queue, buffer_size, chunk_size)
         if starting_values:
             self.worker.data_added.set()
         self.worker.start()
 
     ####################
     # Generator's methods
     ####################
     def __str__(self):
         return f'BufferedGenerator(name={self.name} chunk: {self.chunk_size} on buffer: {self.buffer_size})'
 
     def __iter__(self):
         return self
 
     def __next__(self):
 
         while not self.length and not self.worker.EOF.is_set():
             self.worker.data_added.clear()
             self.worker.data_added.wait(0.2)
 
         if self.length:
             self.worker.keep_reading.set()
             return bytes(self.data_queue.popleft())
         elif self.worker.EOF.is_set():
             raise StopIteration
 
     def throw(self, exc_type, value=None, traceback=None):
         if not self.worker.EOF.is_set():
             raise exc_type(value)
 
     def start(self):
         self.worker.start()
 
     def stop(self):
         self.worker.stop()
 
     def close(self):
         try:
             self.worker.stop()
             self.throw(GeneratorExit)
         except (GeneratorExit, StopIteration):
             pass
 
     ####################
     # Threaded reader's infrastructure.
     ####################
     @property
     def input(self):
         return self.worker.w
 
     @property
     def data_added_event(self):
         return self.worker.data_added
 
     @property
     def data_added(self):
         return self.worker.data_added.is_set()
 
     @property
     def reading_paused(self):
         return not self.worker.keep_reading.is_set()
 
     @property
     def done_reading_event(self):
         """
         Done_reding does not mean that the iterator's buffer is empty.
         Iterator might have done reading from underlying source, but the read
         chunks might still be available for serving through .next() method.
 
         :returns: An Event class instance.
         """
         return self.worker.EOF
 
     @property
     def done_reading(self):
         """
         Done_reading does not mean that the iterator's buffer is empty.
         Iterator might have done reading from underlying source, but the read
         chunks might still be available for serving through .next() method.
 
         :returns: An Bool value.
         """
         return self.worker.EOF.is_set()
 
     @property
     def length(self):
         """
         returns int.
 
         This is the length of the queue of chunks, not the length of
         the combined contents in those chunks.
 
         __len__() cannot be meaningfully implemented because this
         reader is just flying through a bottomless pit content and
         can only know the length of what it already saw.
 
         If __len__() on WSGI server per PEP 3333 returns a value,
         the response's length will be set to that. In order not to
         confuse WSGI PEP3333 servers, we will not implement __len__
         at all.
         """
         return len(self.data_queue)
 
     def prepend(self, x):
         self.data_queue.appendleft(x)
 
     def append(self, x):
         self.data_queue.append(x)
 
     def extend(self, o):
         self.data_queue.extend(o)
 
     def __getitem__(self, i):
         return self.data_queue[i]
 
 
 class SubprocessIOChunker:
     """
     Processor class wrapping handling of subprocess IO.
 
     .. important::
 
        Watch out for the method `__del__` on this class. If this object
        is deleted, it will kill the subprocess, so avoid to
        return the `output` attribute or usage of it like in the following
        example::
 
           # `args` expected to run a program that produces a lot of output
           output = ''.join(SubprocessIOChunker(
              args, shell=False, inputstream=inputstream, env=environ).output)
 
           # `output` will not contain all the data, because the __del__ method
           # has already killed the subprocess in this case before all output
           # has been consumed.
 
 
 
     In a way, this is a "communicate()" replacement with a twist.
 
     - We are multithreaded. Writing in and reading out, err are all sep threads.
     - We support concurrent (in and out) stream processing.
     - The output is not a stream. It's a queue of read string (bytes, not str)
       chunks. The object behaves as an iterable. You can "for chunk in obj:" us.
     - We are non-blocking in more respects than communicate()
       (reading from subprocess out pauses when internal buffer is full, but
        does not block the parent calling code. On the flip side, reading from
        slow-yielding subprocess may block the iteration until data shows up. This
        does not block the parallel inpipe reading occurring parallel thread.)
 
     The purpose of the object is to allow us to wrap subprocess interactions into
     an iterable that can be passed to a WSGI server as the application's return
     value. Because of stream-processing-ability, WSGI does not have to read ALL
     of the subprocess's output and buffer it, before handing it to WSGI server for
     HTTP response. Instead, the class initializer reads just a bit of the stream
     to figure out if error occurred or likely to occur and if not, just hands the
     further iteration over subprocess output to the server for completion of HTTP
     response.
 
     The real or perceived subprocess error is trapped and raised as one of
     OSError family of exceptions
 
     Example usage:
     #    try:
     #        answer = SubprocessIOChunker(
     #            cmd,
     #            input,
     #            buffer_size = 65536,
     #            chunk_size = 4096
     #            )
     #    except (OSError) as e:
     #        print str(e)
     #        raise e
     #
     #    return answer
 
 
     """
 
     # TODO: johbo: This is used to make sure that the open end of the PIPE
     # is closed in the end. It would be way better to wrap this into an
     # object, so that it is closed automatically once it is consumed or
     # something similar.
     _close_input_fd = None
 
     _closed = False
     _stdout = None
     _stderr = None
 
     def __init__(self, cmd, input_stream=None, buffer_size=65536,
                  chunk_size=4096, starting_values=None, fail_on_stderr=True,
                  fail_on_return_code=True, **kwargs):
         """
         Initializes SubprocessIOChunker
 
         :param cmd: A Subprocess.Popen style "cmd". Can be string or array of strings
         :param input_stream: (Default: None) A file-like, string, or file pointer.
         :param buffer_size: (Default: 65536) A size of total buffer per stream in bytes.
         :param chunk_size: (Default: 4096) A max size of a chunk. Actual chunk may be smaller.
         :param starting_values: (Default: []) An array of strings to put in front of output que.
         :param fail_on_stderr: (Default: True) Whether to raise an exception in
                                case something is written to stderr.
         :param fail_on_return_code: (Default: True) Whether to raise an
                                     exception if the return code is not 0.
         """
 
         kwargs['shell'] = kwargs.get('shell', True)
 
         starting_values = starting_values or []
         if input_stream:
             input_streamer = StreamFeeder(input_stream)
             input_streamer.start()
             input_stream = input_streamer.output
             self._close_input_fd = input_stream
 
         self._fail_on_stderr = fail_on_stderr
         self._fail_on_return_code = fail_on_return_code
         self.cmd = cmd
 
         _p = subprocess.Popen(cmd, bufsize=-1, stdin=input_stream, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                               **kwargs)
         self.process = _p
 
         bg_out = BufferedGenerator('stdout', _p.stdout, buffer_size, chunk_size, starting_values)
         bg_err = BufferedGenerator('stderr', _p.stderr, 10240, 1, bottomless=True)
 
         while not bg_out.done_reading and not bg_out.reading_paused and not bg_err.length:
             # doing this until we reach either end of file, or end of buffer.
             bg_out.data_added_event.wait(0.2)
             bg_out.data_added_event.clear()
 
         # at this point it's still ambiguous if we are done reading or just full buffer.
         # Either way, if error (returned by ended process, or implied based on
         # presence of stuff in stderr output) we error out.
         # Else, we are happy.
         return_code = _p.poll()
         ret_code_ok = return_code in [None, 0]
         ret_code_fail = return_code is not None and return_code != 0
+
         if (
             (ret_code_fail and fail_on_return_code) or
             (ret_code_ok and fail_on_stderr and bg_err.length)
         ):
 
             try:
                 _p.terminate()
             except Exception:
                 pass
 
             bg_out.stop()
             out = b''.join(bg_out)
             self._stdout = out
 
             bg_err.stop()
             err = b''.join(bg_err)
             self._stderr = err
 
             # code from https://github.com/schacon/grack/pull/7
             if err.strip() == b'fatal: The remote end hung up unexpectedly' and out.startswith(b'0034shallow '):
                 bg_out = iter([out])
                 _p = None
             elif err and fail_on_stderr:
                 text_err = err.decode()
                 raise OSError(
                     f"Subprocess exited due to an error:\n{text_err}")
 
             if ret_code_fail and fail_on_return_code:
                 text_err = err.decode()
                 if not err:
                     # maybe get empty stderr, try stdout instead
                     # in many cases git reports the errors on stdout too
                     text_err = out.decode()
                 raise OSError(
                     f"Subprocess exited with non 0 ret code:{return_code}: stderr:{text_err}")
 
         self.stdout = bg_out
         self.stderr = bg_err
         self.inputstream = input_stream
 
     def __str__(self):
         proc = getattr(self, 'process', 'NO_PROCESS')
         return f'SubprocessIOChunker: {proc}'
 
     def __iter__(self):
         return self
 
     def __next__(self):
         # Note: mikhail: We need to be sure that we are checking the return
         # code after the stdout stream is closed. Some processes, e.g. git
         # are doing some magic in between closing stdout and terminating the
         # process and, as a result,  we are not getting return code on "slow"
         # systems.
         result = None
         stop_iteration = None
         try:
             result = next(self.stdout)
         except StopIteration as e:
             stop_iteration = e
 
         if self.process:
             return_code = self.process.poll()
             ret_code_fail = return_code is not None and return_code != 0
             if ret_code_fail and self._fail_on_return_code:
                 self.stop_streams()
                 err = self.get_stderr()
                 raise OSError(
                     f"Subprocess exited (exit_code:{return_code}) due to an error during iteration:\n{err}")
 
         if stop_iteration:
             raise stop_iteration
         return result
 
     def throw(self, exc_type, value=None, traceback=None):
         if self.stdout.length or not self.stdout.done_reading:
             raise exc_type(value)
 
     def close(self):
         if self._closed:
             return
 
         try:
             self.process.terminate()
         except Exception:
             pass
         if self._close_input_fd:
             os.close(self._close_input_fd)
         try:
             self.stdout.close()
         except Exception:
             pass
         try:
             self.stderr.close()
         except Exception:
             pass
         try:
             os.close(self.inputstream)
         except Exception:
             pass
 
         self._closed = True
 
     def stop_streams(self):
         getattr(self.stdout, 'stop', lambda: None)()
         getattr(self.stderr, 'stop', lambda: None)()
 
     def get_stdout(self):
         if self._stdout:
             return self._stdout
         else:
             return b''.join(self.stdout)
 
     def get_stderr(self):
         if self._stderr:
             return self._stderr
         else:
             return b''.join(self.stderr)
 
 
 def run_command(arguments, env=None):
     """
     Run the specified command and return the stdout.
 
     :param arguments: sequence of program arguments (including the program name)
     :type arguments: list[str]
     """
 
     cmd = arguments
     log.debug('Running subprocessio command %s', cmd)
     proc = None
     try:
         _opts = {'shell': False, 'fail_on_stderr': False}
         if env:
             _opts.update({'env': env})
         proc = SubprocessIOChunker(cmd, **_opts)
         return b''.join(proc), b''.join(proc.stderr)
     except OSError as err:
         cmd = ' '.join(map(safe_str, cmd))  # human friendly CMD
         tb_err = ("Couldn't run subprocessio command (%s).\n"
                   "Original error was:%s\n" % (cmd, err))
         log.exception(tb_err)
         raise Exception(tb_err)
     finally:
         if proc:
             proc.close()