Files @ ee1da602b590
Branch filter:

Location: kallithea/init.d/kallithea-upstart.conf

Thomas De Schampheleire
repo_groups: fix deletion of subgroups

Deletion of a repository group that has a parent group (i.e. is not at the
root of the repository group tree) failed as follows:

Traceback (most recent call last):
[...]
File ".../lib/python3.9/site-packages/tg/configurator/components/dispatch.py", line 114, in _call_controller
return controller(*remainder, **params)
File "<decorator-gen-5>", line 2, in delete

File "/home/tdescham/repo/contrib/kallithea/kallithea-release/kallithea/lib/auth.py", line 572, in __wrapper
return func(*fargs, **fkwargs)
File "/home/tdescham/repo/contrib/kallithea/kallithea-release/kallithea/controllers/admin/repo_groups.py", line 271, in delete
if gr.parent_group:
File ".../lib/python3.9/site-packages/sqlalchemy/orm/attributes.py", line 294, in __get__
return self.impl.get(instance_state(instance), dict_)
File ".../lib/python3.9/site-packages/sqlalchemy/orm/attributes.py", line 730, in get
value = self.callable_(state, passive)
File ".../lib/python3.9/site-packages/sqlalchemy/orm/strategies.py", line 717, in _load_for_state
raise orm_exc.DetachedInstanceError(
sqlalchemy.orm.exc.DetachedInstanceError: Parent instance <RepoGroup at 0x7f1f2664f4c0> is not bound to a Session; lazy load operation of attribute 'parent_group' cannot proceed (Background on this error at: http://sqlalche.me/e/13/bhk3)


In the reference 'gr.parent_group', 'gr' is an SQLAlchemy object referring
to the group being deleted, and 'gr.parent_group' is a lazy reference to its
parent group. The 'lazy' means that the parent group object is not loaded
automatically when 'gr' is assigned, but instead will be loaded on-the-fly
when the parent group is actually accessed. See [1] and [2] for more
information.

The problem was that the lazy 'parent_group' attribute was accessed _after_
deleting the database object it was part of.

Fix this by obtaining a handle to the parent group _before_ deleting the
subgroup.

Reported-by: André Klitzing (via mailing list)


[1] https://docs.sqlalchemy.org/en/13/errors.html#error-bhk3
[2] https://docs.sqlalchemy.org/en/13/orm/loading_relationships.html
# kallithea - run the kallithea daemon as an upstart job
# Change variables/paths as necessary and place file /etc/init/kallithea.conf
# start/stop/restart as normal upstart job (ie: $ start kallithea)

description     "Kallithea Mercurial Server"
author          "Matt Zuba <matt.zuba@goodwillaz.org"

start on (local-filesystems and runlevel [2345])
stop on runlevel [!2345]

respawn

umask 0022

env PIDFILE=/var/hg/kallithea/kallithea.pid
env LOGFILE=/var/hg/kallithea/log/kallithea.log
env APPINI=/var/hg/kallithea/production.ini
env HOME=/var/hg
env USER=hg
env GROUP=hg

exec /var/hg/.virtualenvs/kallithea/bin/gearbox serve --user=$USER --group=$GROUP --pid-file=$PIDFILE --log-file=$LOGFILE -c $APPINI

post-stop script
    rm -f $PIDFILE
end script