Files
@ ee1da602b590
Branch filter:
Location: kallithea/scripts/dbmigrate-test - annotation
ee1da602b590
3.5 KiB
text/plain
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
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
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 | 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b f973b866fffc 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 3158cf0dafb7 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b 51b1af58589b | #!/bin/sh -e
if [ $# -lt 2 ] || [ $# -gt 3 ]; then
cat >&2 <<EOD
usage: $0 CONFIG_FILE FROM_REV [TO_REV]
Runs a database migration from FROM_REV to TO_REV (default: current
working directory parent), using the specified CONFIG_FILE (.ini file).
Test is run using a clean Kallithea install, in a temporary virtual
environment. FROM_REV and (optional) TO_REV should be Mercurial revision
identifiers (e.g. changeset hash or a version number tag). The working
directory is not touched, but the database referenced in the config file
will be (re)created.
Only SQLite is available out of the box; for MySQL or PostgreSQL, set
the EXTRA environment variable to the required package(s), and it'll
be installed in the virtual environment. (E.g. EXTRA=MySQL-python or
EXTRA=psycopg2.)
The temporary directory is not removed, allowing follow-up examination
of the upgrade results. It is, however, created in /tmp by default,
which many Linux distributions automatically clean at regular intervals.
EOD
exit 1
fi
config_file=$(readlink -f "$1")
from_rev=$2
to_rev=$3
source_repo=$(dirname "$(dirname "$(readlink -f "$0")")")
announce() {
echo
echo "$1"
echo
}
quiet_if_ok() (
local output
local st
set +e
output=$("$@" < /dev/null 2>&1)
st=$?
if [ $st -ne 0 ]; then
echo "$output" >&2
echo "Command $@ returned exit status $st." >&2
exit 1
fi
)
HG() {
"${HG:-hg}" --repository "$source_repo" "$@"
}
# If upgrading to "current revision", warn if working directory is dirty.
if [ ! "$to_rev" ] && [ "$(HG status -mard)" ]; then
announce "Warning: Uncommitted changes in working directory will be ignored!"
fi
from_rev_hash=$(HG id --id --rev "${from_rev:-.}")
to_rev_hash=$(HG id --id --rev "${to_rev:-.}")
temp=$(readlink -f "$(mktemp --tmpdir -d 'dbmigrate-test.XXXXXX')")
cat <<EOD
Config file: $config_file
EOD
sed -n -e 's/^sqlalchemy\.url *= */Database URL: /p' "$config_file"
cat <<EOD
Working dir: $temp
Repository: $source_repo
Upgrade from: $from_rev_hash (${from_rev:-current})
Upgrade to: $to_rev_hash (${to_rev:-current})
Extra packages: ${EXTRA:-(none)}
EOD
mkdir "$temp/repos" # empty
# Enable caching for old pip versions (this will cache the pip upgrade)
# Newer pip versions cache automatically, and don't use this variable.
if [ ! "$PIP_DOWNLOAD_CACHE" ]; then
export PIP_DOWNLOAD_CACHE=$HOME/.cache/pip/legacy
fi
install_kallithea() {
local prefix=$1
local rev=$2
announce "Installing Kallithea $rev in $prefix..."
"${VIRTUALENV:-virtualenv}" --quiet "$prefix-env"
HG archive --rev "$rev" "$prefix"
(
cd "$prefix"
. "$prefix-env/bin/activate"
pip install --quiet --upgrade pip setuptools mercurial $EXTRA
pip install --quiet -e .
)
}
install_kallithea "$temp/from" "$from_rev_hash"
(
cd "$temp/from"
. "$temp/from-env/bin/activate"
announce "Initializing database..."
quiet_if_ok kallithea-cli db-create -c "$config_file" --repos="$temp/repos" --user=doe --email=doe@example.com --password=123456 --no-public-access --force-yes
alembic -c "$config_file" current -v
)
install_kallithea "$temp/to" "$to_rev_hash"
(
cd "$temp/to"
. "$temp/to-env/bin/activate"
announce "Commencing database upgrade from shown Alembic revision to head..."
alembic -c "$config_file" current -v
alembic -c "$config_file" upgrade head
announce "Upgrade complete, now at the shown Alembic revision:"
alembic -c "$config_file" current -v
)
|