flat-manager-client

#!/usr/bin/python3

# 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 2 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 asyncio
import base64
import binascii
import errno
import fnmatch
import gzip
import json
import logging
import os
import sys
import time
import traceback
from argparse import ArgumentParser
from functools import reduce
from urllib.parse import urljoin, urlparse, urlsplit, urlunparse, urlunsplit

import aiohttp
from tenacity import retry, retry_if_exception_type, stop_after_attempt, wait_fixed

import gi
gi.require_version('OSTree', '1.0')
from gi.repository import Gio, GLib, OSTree

UPLOAD_CHUNK_LIMIT = 4 * 1024 * 1024
DEFAULT_LIMIT = 2 ** 16

def eprint(*args, **kwargs):
    print(*args, file=sys.stderr, **kwargs)

class UsageException(Exception):
    def __init__(self, msg):
        self.msg = msg
    def __str__(self):
        return self.msg

class ApiError(Exception):
    def __init__(self, response, body):
        self.url = str(response.url)
        self.status = response.status

        try:
            self.body = json.loads(response);
        except:
            self.body = {"status": self.status, "error-type": "no-error", "message": "No json error details from server"}

    def repr(self):
        return {
            "type": "api",
            "url": self.url,
            "status_code": self.status,
            "details": self.body
        }

    def __str__(self):
        return "Api call to %s failed with status %d, details: %s" % (self.url, self.status, self.body)


# This is similar to the regular payload, but opens the file lazily
class AsyncNamedFilePart(aiohttp.payload.Payload):
    def __init__(self,
                 value,
                 disposition='attachment',
                 *args,
                 **kwargs):
        self._file = None
        if 'filename' not in kwargs:
            kwargs['filename'] = os.path.basename(value)

        super().__init__(value, *args, **kwargs)

        if self._filename is not None and disposition is not None:
            self.set_content_disposition(disposition, filename=self._filename, quote_fields=False)

        self._size = os.stat(value).st_size

    async def write(self, writer):
        if self._file is None or self._file.closed:
            self._file = open(self._value, 'rb')
        try:
            chunk = self._file.read(DEFAULT_LIMIT)
            while chunk:
                await writer.write(chunk)
                chunk = self._file.read(DEFAULT_LIMIT)
        finally:
            self._file.close()

    @property
    def size(self):
        return self._size

def ostree_object_path(repo, obj):
    repodir = repo.get_path().get_path()
    return os.path.join(repodir, 'objects', obj[0:2], obj[2:])

def ostree_get_dir_files(repo, objects, dirtree):
    if dirtree.endswith(".dirtree"):
        dirtree = dirtree[:-8]
    dirtreev = repo.load_variant(OSTree.ObjectType.DIR_TREE, dirtree)[1]
    iter = OSTree.RepoCommitTraverseIter()
    iter.init_dirtree(repo, dirtreev, 0)
    while True:
        type = iter.next()
        if type == OSTree.RepoCommitIterResult.END:
            break
        if type == OSTree.RepoCommitIterResult.ERROR:
            break
        if type == OSTree.RepoCommitIterResult.FILE:
            d = iter.get_file()
            objects.add(d.out_checksum + ".filez")
        if type == OSTree.RepoCommitIterResult.DIR:
            pass

def local_needed_files(repo, metadata_objects):
    objects = set()
    for c in metadata_objects:
        if c.endswith(".dirtree"):
            ostree_get_dir_files(repo, objects, c)
    return objects

def local_needed_metadata_dirtree(repo, objects, dirtree_content, dirtree_meta):
    objects.add(dirtree_meta + ".dirmeta")
    dirtree_content_name = dirtree_content + ".dirtree"
    if dirtree_content_name in objects:
        return
    objects.add(dirtree_content_name)

    dirtreev = repo.load_variant(OSTree.ObjectType.DIR_TREE, dirtree_content)[1]
    iter = OSTree.RepoCommitTraverseIter()
    iter.init_dirtree(repo, dirtreev, 0)
    while True:
        type = iter.next()
        if type == OSTree.RepoCommitIterResult.END:
            break
        if type == OSTree.RepoCommitIterResult.ERROR:
            break
        if type == OSTree.RepoCommitIterResult.FILE:
            pass
        if type == OSTree.RepoCommitIterResult.DIR:
            d = iter.get_dir()
            local_needed_metadata_dirtree(repo, objects, d.out_content_checksum, d.out_meta_checksum)

def local_needed_metadata(repo, commits):
    objects = set()
    for rev in commits:
        objects.add(rev + ".commit")
        commitv = repo.load_variant(OSTree.ObjectType.COMMIT, rev)[1]
        iter = OSTree.RepoCommitTraverseIter()
        iter.init_commit(repo, commitv, 0)
        while True:
            type = iter.next()
            if type == OSTree.RepoCommitIterResult.END:
                break
            if type == OSTree.RepoCommitIterResult.ERROR:
                break
            if type == OSTree.RepoCommitIterResult.FILE:
                pass
            if type == OSTree.RepoCommitIterResult.DIR:
                d = iter.get_dir()
                local_needed_metadata_dirtree(repo, objects, d.out_content_checksum, d.out_meta_checksum)
    return objects


def chunks(l, n):
    """Yield successive n-sized chunks from l."""
    for i in range(0, len(l), n):
        yield l[i:i + n]

async def missing_objects(session, build_url, token, wanted):
    missing=[]
    for chunk in chunks(wanted, 2000):
        wanted_json=json.dumps({'wanted': chunk}).encode('utf-8')
        data=gzip.compress(wanted_json)
        headers = {
            'Authorization': 'Bearer ' + token,
            'Content-Encoding': 'gzip',
            'Content-Type': 'application/json'
        }
        resp = await session.get(build_url + "/missing_objects", data=data, headers=headers)
        async with resp:
            if resp.status != 200:
                raise ApiError(resp, await resp.text())
            data = await resp.json()
            missing.extend(data["missing"])
    return missing

async def upload_files(session, build_url, token, files):
    if len(files) == 0:
        return
    print("Uploading %d files (%d bytes)" % (len(files), reduce(lambda x, y: x + y, map(lambda f: f.size, files))))
    with aiohttp.MultipartWriter() as writer:
        for f in files:
            writer.append(f)
    writer.headers['Authorization'] = 'Bearer ' + token
    resp = await session.request("post", build_url + '/upload', data=writer, headers=writer.headers)
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())

async def upload_deltas(session, repo_path, build_url, token, deltas, refs, ignore_delta):
    if not len(deltas):
        return

    req = []
    for ref, commit in refs.items():
        # Skip screenshots here
        parts = ref.split("/")
        if len(parts) == 4 and (parts[0] == "app" or parts[0] =="runtime") and not should_skip_delta(parts[1], ignore_delta):
            for delta in deltas:
                # Only upload from-scratch deltas, as these are the only reused ones
                if delta == commit:
                    print(" %s: %s" % (ref, delta))
                    delta_name = delta_name_encode (delta)
                    delta_dir = repo_path + "/deltas/" + delta_name[:2] + "/" + delta_name[2:]
                    parts = os.listdir(delta_dir)
                    for part in parts:
                        req.append(AsyncNamedFilePart(delta_dir + "/" + part, filename = delta_name + "." + part + ".delta"))

    if len(req):
        await upload_files(session, build_url, token, req)


async def upload_objects(session, repo_path, build_url, token, objects):
    req = []
    total_size = 0
    for file_obj in objects:
        named = get_object_multipart(repo_path, file_obj)
        file_size = named.size
        if total_size + file_size > UPLOAD_CHUNK_LIMIT: # The new object would bring us over the chunk limit
            if len(req) > 0: # We already have some objects, upload those first
                next_req = [named]
                total_size = file_size
            else:
                next_req  =  []
                req.append(named)
                total_size = 0
            await upload_files(session, build_url, token, req)
            req = next_req
        else:
            total_size = total_size + file_size
            req.append(named)

    # Upload any remainder
    await upload_files(session, build_url, token, req)

async def create_ref(session, build_url, token, ref, commit):
    print("Creating ref %s with commit %s" % (ref, commit))
    resp = await session.post(build_url + "/build_ref", headers={'Authorization': 'Bearer ' + token}, json= { "ref": ref, "commit": commit} )
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())

        data = await resp.json()
        return data

async def add_extra_ids(session, build_url, token, extra_ids):
    print("Adding extra ids %s" % (extra_ids))
    resp = await session.post(build_url + "/add_extra_ids", headers={'Authorization': 'Bearer ' + token}, json= { "ids": extra_ids} )
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())

        data = await resp.json()
        return data

async def get_build(session, build_url, token):
    resp = await session.get(build_url, headers={'Authorization': 'Bearer ' + token})
    if resp.status != 200:
        raise ApiError(resp, await resp.text())
    data = await resp.json()
    return data

# For stupid reasons this is a string with json, lets expand it
def reparse_job_results(job):
    job["results"] = json.loads(job.get("results", "{}"))
    return job

async def get_job(session, job_url, token):
    resp = await session.get(job_url, headers={'Authorization': 'Bearer ' + token}, json={})
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())
        data = await resp.json()
        return data

async def wait_for_job(session, job_url, token):
    reported_delay = False
    old_job_status  = 0
    printed_len = 0
    iterations_since_change=0
    error_iterations = 0
    while True:
        try:
            resp = await session.get(job_url, headers={'Authorization': 'Bearer ' + token}, json={'log-offset': printed_len})
            async with resp:
                if resp.status == 200:
                    error_iterations = 0
                    job = await resp.json()
                    job_status = job['status']
                    if job_status == 0 and not reported_delay:
                        reported_delay = True
                        start_after_struct = job.get("start_after", None)
                        if start_after_struct:
                            start_after = start_after_struct.get("secs_since_epoch", None)
                            now = time.time()
                            if start_after and start_after > now:
                                print("Waiting %d seconds before starting job" % (int(start_after - now)))
                    if job_status > 0 and old_job_status == 0:
                        print("/ Job was started");
                    old_job_status = job_status
                    log = job['log']
                    if len(log) > 0:
                        iterations_since_change=0
                        for line in log.splitlines(True):
                            print("| %s" % line, end="")
                        printed_len = printed_len + len(log)
                    else:
                        iterations_since_change=iterations_since_change+1
                    if job_status > 1:
                        if job_status == 2:
                            print("\ Job completed successfully")
                        else:
                            print("\ Job failed")
                        return job
                else:
                    iterations_since_change=4 # Start at 4 so we ramp up the delay faster
                    error_iterations=error_iterations + 1
                    if error_iterations <= 5:
                        print("Unexpected response %s getting job log, ignoring" % resp.status)
                    else:
                        raise ApiError(resp, await resp.text())
        except OSError as e:
            if e.args[0] == errno.ECONNRESET:
                # Client disconnected, retry
                # Not sure exactly why, but i got a lot of ConnectionResetErrors here
                # in tests. I guess the server stops reusing a http2 session after a bit
                # Should be fine to retry with the backof
                pass
            else:
                raise
        # Some polling backoff to avoid loading the server
        if iterations_since_change <= 1:
            sleep_time=1
        elif iterations_since_change < 5:
            sleep_time=3
        elif iterations_since_change < 15:
            sleep_time=5
        elif iterations_since_change < 30:
            sleep_time=10
        else:
            sleep_time=60
        time.sleep(sleep_time)

async def commit_build(session, build_url, eol, eol_rebase, token_type, wait, token):
    print("Committing build %s" % (build_url))
    json = {
        "endoflife": eol,
        "endoflife_rebase": eol_rebase
    }
    if token_type != None:
        json['token_type'] = token_type
    resp = await session.post(build_url + "/commit", headers={'Authorization': 'Bearer ' + token}, json=json)
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())

        job = await resp.json()
        job_url = resp.headers['location'];

        if wait:
            print("Waiting for commit job")
            job = await wait_for_job(session, job_url, token);

        reparse_job_results(job)
        job["location"] = job_url
        return job

async def publish_build(session, build_url, wait, token):
    print("Publishing build %s" % (build_url))
    resp = await session.post(build_url + "/publish", headers={'Authorization': 'Bearer ' + token}, json= { } )
    async with resp:
        if resp.status == 400:
            body = await resp.text()
            try:
                msg = json.loads(body)
                if msg.get("current-state", "") == "published":
                    print("the build has been already published")
                    return {}
            except:
                pass

        if resp.status != 200:
            raise ApiError(resp, await resp.text())

        job = await resp.json()
        job_url = resp.headers['location'];

        if wait:
            print("Waiting for publish job")
            job = await wait_for_job(session, job_url, token);

        reparse_job_results(job)
        job["location"] = job_url
        return job

async def purge_build(session, build_url, token):
    print("Purging build %s" % (build_url))
    resp = await session.post(build_url + "/purge", headers={'Authorization': 'Bearer ' + token}, json= {} )
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())
        return await resp.json()

async def create_token(session, manager_url, token, name, subject, scope, duration):
    token_url = urljoin(manager_url, "api/v1/token_subset")
    resp = await session.post(token_url, headers={'Authorization': 'Bearer ' + token}, json = {
        "name": name,
        "sub": subject,
        "scope": scope,
        "duration": duration,
    })
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())
        return await resp.json()

def get_object_multipart(repo_path, object):
    return AsyncNamedFilePart(repo_path + "/objects/" + object[:2] + "/" + object[2:], filename=object)

async def create_command(session, args):
    build_url = urljoin(args.manager_url, "api/v1/build")
    resp = await session.post(build_url, headers={'Authorization': 'Bearer ' + args.token}, json={
        "repo": args.repo
    })
    async with resp:
        if resp.status != 200:
            raise ApiError(resp, await resp.text())
        data = await resp.json()
        data["location"] = resp.headers['location']
        if not args.print_output:
            print(resp.headers['location'])
        return data

def delta_name_part_encode(commit):
    return base64.b64encode(binascii.unhexlify(commit), b"+_")[:-1].decode("utf-8")

def delta_name_encode (delta):
    return "-".join(map(delta_name_part_encode, delta.split("-")))

def should_skip_delta(id, globs):
    if globs:
        for glob in globs:
            if fnmatch.fnmatch(id, glob):
                return True
    return False

def build_url_to_api(build_url):
    parts = urlparse(build_url)
    path = os.path.dirname(os.path.dirname(parts.path))
    return urlunparse((parts.scheme, parts.netloc, path, None, None, None))

@retry(
    stop=stop_after_attempt(6),
    wait=wait_fixed(10),
    retry=retry_if_exception_type(ApiError),
    reraise=True,
)
async def push_command(session, args):
    local_repo = OSTree.Repo.new(Gio.File.new_for_path(args.repo_path))
    try:
        local_repo.open(None)
    except GLib.Error as err:
        raise UsageException("Can't open repo %s: %s" % (args.repo_path, err.message)) from err

    refs = {}
    if len(args.branches) == 0:
        _, all_refs = local_repo.list_refs(None, None)
        for ref in all_refs:
            if ref.startswith("app/") or ref.startswith("runtime/") or ref.startswith("screenshots/"):
                refs[ref] = all_refs[ref]
    else:
        for branch in args.branches:
            _, rev = local_repo.resolve_rev(branch, False)
            refs[branch] = rev

    if (args.minimal_token):
        id = os.path.basename(urlparse(args.build_url).path)
        token = create_token(args.build_url, args.token, "minimal-upload", "build/%s" % (id), ["upload"], 60*60)["token"]
    else:
        token = args.token

    print("Uploading refs to %s: %s"% (args.build_url, list(refs)))

    metadata_objects = local_needed_metadata(local_repo, refs.values())

    print("Refs contain %d metadata objects" % (len(metadata_objects)))

    missing_metadata_objects = await missing_objects(session, args.build_url, token, list(metadata_objects))

    print("Remote missing %d of those" % (len(missing_metadata_objects)))

    file_objects = local_needed_files(local_repo, missing_metadata_objects)
    print("Has %d file objects for those" % (len(file_objects)))

    missing_file_objects = await missing_objects(session, args.build_url, token, list(file_objects))
    print("Remote missing %d of those" % (len(missing_file_objects)))

    # First upload all missing file objects
    print("Uploading file objects")
    await upload_objects(session, args.repo_path, args.build_url, token, missing_file_objects)

    # Then all the metadata
    print("Uploading metadata objects")
    await upload_objects(session, args.repo_path, args.build_url, token, missing_metadata_objects)

    _, deltas = local_repo.list_static_delta_names()
    print("Uploading deltas")
    await upload_deltas(session, args.repo_path, args.build_url, token, deltas, refs, args.ignore_delta)

    # Then the refs
    for ref, commit in refs.items():
        await create_ref(session, args.build_url, token, ref, commit)

    # Then any extra ids
    if args.extra_id:
        await add_extra_ids(session, args.build_url, token, args.extra_id)

    commit_job = None
    publish_job = None
    update_job = None

    # Note, this always uses the full token, as the minimal one only has upload permissions
    if args.commit or args.publish:
        commit_job = await commit_build(session, args.build_url, args.end_of_life, args.end_of_life_rebase, args.token_type, args.publish or args.wait, args.token)

    if args.publish:
        publish_job = await publish_build(session, args.build_url, args.wait or args.wait_update, args.token)
        update_job_id = publish_job.get("results", {}).get("update-repo-job", None)
        if update_job_id:
            print("Queued repo update job %d" %(update_job_id))
            update_job_url = build_url_to_api(args.build_url) + "/job/" + str(update_job_id)
            if args.wait_update:
                print("Waiting for repo update job")
                update_job = await wait_for_job (session, update_job_url, token);
            else:
                update_job = await get_job(session, update_job_url, token)
            reparse_job_results(update_job)
            update_job["location"] = update_job_url

    data = await get_build(session, args.build_url, args.token)
    if commit_job:
        data["commit_job"] = commit_job
    if publish_job:
        data["publish_job"] = publish_job
    if update_job:
        data["update_job"] = update_job
    return data

async def commit_command(session, args):
    job = await commit_build(session, args.build_url, args.end_of_life, args.end_of_life_rebase, args.token_type, args.wait, args.token)
    return job

async def publish_command(session, args):
    job = await publish_build(session, args.build_url, args.wait or args.wait_update, args.token)
    update_job_id = job.get("results", {}).get("update-repo-job", None)
    if update_job_id:
        print("Queued repo update job %d" %(update_job_id))
        update_job_url = build_url_to_api(args.build_url) + "/job/" + str(update_job_id)
        if args.wait_update:
            print("Waiting for repo update job")
            update_job = await wait_for_job(session, update_job_url, args.token);
        else:
            update_job = await get_job(session, update_job_url, args.token)
        reparse_job_results(update_job)
        update_job["location"] = update_job_url
    return job

async def purge_command(session, args):
    job = await purge_build(session, args.build_url, args.token)
    return job

async def create_token_command(session, args):
    data = await create_token(session, args.manager_url, args.token, args.name, args.subject, args.scope, args.duration)
    if not args.print_output:
        print(data['token'])
    return data

async def follow_job_command(session, args):
    job = await wait_for_job(session, args.job_url, args.token)
    return job

async def run_with_session(args):
    timeout = aiohttp.ClientTimeout(total=90*60)
    async with aiohttp.ClientSession(timeout=timeout) as session:
        result = await args.func(session, args)
    return result

if __name__ == '__main__':
    progname = os.path.basename(sys.argv[0])

    parser = ArgumentParser(prog=progname)
    parser.add_argument('-v', '--verbose', action='store_true',
                        help='enable verbose output')
    parser.add_argument('--debug', action='store_true',
                        help='enable debugging output')
    parser.add_argument('--output', help='Write output json to file')
    parser.add_argument('--print-output', action='store_true', help='Print output json')
    parser.add_argument('--token', help='use this token')
    parser.add_argument('--token-file', help='use token from file')
    subparsers = parser.add_subparsers(title='subcommands',
                                       dest='subparser_name',
                                       description='valid subcommands',
                                       help='additional help')

    create_parser = subparsers.add_parser('create', help='Create new build')
    create_parser.add_argument('manager_url', help='remote repo manager url')
    create_parser.add_argument('repo', help='repo name')
    create_parser.set_defaults(func=create_command)

    push_parser = subparsers.add_parser('push', help='Push to repo manager')
    push_parser.add_argument('build_url', help='remote build url')
    push_parser.add_argument('repo_path', help='local repository')
    push_parser.add_argument('branches', nargs='*', help='branches to push')
    push_parser.add_argument('--commit', action='store_true',
                             help='commit build after pushing')
    push_parser.add_argument('--publish', action='store_true',
                             help='publish build after committing')
    push_parser.add_argument('--extra-id', action='append', help='add extra collection-id')
    push_parser.add_argument('--ignore-delta', action='append', help='don\'t upload deltas matching this glob')
    push_parser.add_argument('--wait', action='store_true',
                             help='wait for commit/publish to finish')
    push_parser.add_argument('--wait-update', action='store_true',
                             help='wait for update-repo to finish')
    push_parser.add_argument('--minimal-token', action='store_true',
                             help='Create minimal token for the upload')
    push_parser.add_argument('--end-of-life', help='Set end of life')
    push_parser.add_argument('--end-of-life-rebase', help='Set new ID which will supercede the current one')
    push_parser.add_argument('--token-type', help='Set token type', type=int)
    push_parser.set_defaults(func=push_command)

    commit_parser = subparsers.add_parser('commit', help='Commit build')
    commit_parser.add_argument('--wait', action='store_true',
                             help='wait for commit to finish')
    commit_parser.add_argument('--end-of-life', help='Set end of life')
    commit_parser.add_argument('--end-of-life-rebase', help='Set new ID which will supercede the current one')
    commit_parser.add_argument('--token-type', help='Set token type', type=int)
    commit_parser.add_argument('build_url', help='remote build url')
    commit_parser.set_defaults(func=commit_command)

    publish_parser = subparsers.add_parser('publish', help='Publish build')
    publish_parser.add_argument('--wait', action='store_true',
                             help='wait for publish to finish')
    publish_parser.add_argument('--wait-update', action='store_true',
                             help='wait for update-repo to finish')
    publish_parser.add_argument('build_url', help='remote build url')
    publish_parser.set_defaults(func=publish_command)

    purge_parser = subparsers.add_parser('purge', help='Purge build')
    purge_parser.add_argument('build_url', help='remote build url')
    purge_parser.set_defaults(func=purge_command)

    create_token_parser = subparsers.add_parser('create-token', help='Create subset token')
    create_token_parser.add_argument('manager_url', help='remote repo manager url')
    create_token_parser.add_argument('name', help='Name')
    create_token_parser.add_argument('subject', help='Subject')
    create_token_parser.add_argument('scope', nargs='*', help='Scope')
    create_token_parser.add_argument('--duration', help='Duration until expires, in seconds',
                                     default=60*60*24, # Default duration is one day
                                     type=int)
    create_token_parser.set_defaults(func=create_token_command)

    follow_job_parser = subparsers.add_parser('follow-job', help='Follow existing job log')
    follow_job_parser.add_argument('job_url', help='url of job')
    follow_job_parser.set_defaults(func=follow_job_command)

    args = parser.parse_args()

    loglevel = logging.WARNING
    if args.verbose:
        loglevel = logging.INFO
    if args.debug:
        loglevel = logging.DEBUG

    logging.basicConfig(format='%(module)s: %(levelname)s: %(message)s',
                        level=loglevel, stream=sys.stderr)

    if not args.subparser_name:
        print("No subcommand specified, see --help for usage")
        exit(1)

    if not args.token:
        if args.token_file:
            file = open(args.token_file, 'rb')
            args.token = file.read().splitlines()[0].decode("utf-8").strip()
        elif "REPO_TOKEN" in os.environ:
            args.token = os.environ["REPO_TOKEN"]
        else:
            print("No token available, pass with --token, --token-file or $REPO_TOKEN")
            exit(1)


    res = 1
    output = None
    try:
        loop = asyncio.get_event_loop()
        result = loop.run_until_complete(run_with_session(args))

        output = {
            "command": args.subparser_name,
            "result": result,
        }
        res = 0
    except SystemExit:
        # Something called sys.exit(), lets just exit
        res = 1
        raise # Pass on regular exit callse
    except ApiError as e:
        eprint(str(e))
        output = {
            "command": args.subparser_name,
            "error": e.repr(),
        }
    except UsageException as e:
        eprint(str(e))
        output = {
            "error": {
                "type": "usage",
                "details": {
                    "message": str(e),
                }
            }
        }
    except:
        ei = sys.exc_info()
        eprint("Unexpected %s exception in %s: %s" % (ei[0].__name__, args.subparser_name, ei[1]))
        eprint(traceback.format_exc())
        output = {
            "command": args.subparser_name,
            "error": {
                "type": "exception",
                "details": {
                    "error-type": ei[0].__name__,
                    "message": str(ei[1]),
                }
            }
        }
        res = 1

    if output:
        if args.print_output:
            print(json.dumps(output, indent=4))
        if args.output:
            f = open(args.output,"w+")
            f.write(json.dumps(output, indent=4))
            f.write("\n")
            f.close()
    exit(res)