- As of bind 9.16, the option to enable DNSSEC 'dnssec-enable' is obsolete and
has no effect[1]. The option 'dnssec-validation' controls DNSSEC validation and
is set to 'auto' by default. 'auto' means that DNSSEC validation is enabled and
default trust anchor is used for DNS root zone. DNSSEC signatures are also
passed onto a client whenever available. Current stable, Debian Buster, has
version 9.16[3].
- As of bind 9.18, the option to enable DNSSEC 'dnssec-enable' is not recognized
and causes the daemon to fail to start[2]. Debian next, Debian Bookworm, has
version 9.18[3]. Therefore, in testing and unstable, bind fails to start of
installation from FreedomBox.
- There is no use-case for changing the current default behavior.
Links:
1)
https://bind9.readthedocs.io/en/v9_16_32/reference.html#dnssec-validation-option
2) https://bind9.readthedocs.io/en/v9_18_6/reference.html
3) https://tracker.debian.org/pkg/bind9
Tests:
- Run functional and unit tests.
- Option to enable/disable DNSSEC is removed.
- When bind is installed on testing without the patch, it fails to start. When
the patch is applied, bind will be upgraded, the dnssec-enable option is removed
from the configuration file /etc/bind/named.conf.options and bind is running.
Signed-off-by: Sunil Mohan Adapa <sunil@medhas.org>
Reviewed-by: James Valleroy <jvalleroy@mailbox.org>
- ugettext functions will be removed in Django 4.0. Each use emits a warning
when running with Django 3.2. Since we have warnings enabled in developer mode,
we see quite a few messages because of this.
- ugettext is already a simple alias of gettext. So, no regressions are
expected.
Tests:
- Accessing an affected app in UI with Django 3.2 and Django 2.2 works fine.
- Using Django 3.2 there are no warnings related to removal of ugettext
functions.
- Ran regular unit tests.
Signed-off-by: Sunil Mohan Adapa <sunil@medhas.org>
Reviewed-by: James Valleroy <jvalleroy@mailbox.org>
- Introduce new API to mark an app that it can't be disabled.
- Mark jsxc, storage, config, upgrade and firewall apps as can't be disabled.
- Fixed functional tests
- Replaced AppForm with forms.Form in all modules' forms.py.
- Remove app.template.js.
- Remove unused styles.
- Remove app status checks in form_valid of Deluge, Diaspora, Matrix, Ejabberd,
MediaWiki, Storage, Transmission, Quassel
- Purge unused is_enabled context variables (Ikiwiki)
- ejabberd: Minor cleanup in template
- jsxc: Cleanup unneeded overrides
- tahoe: Cleanup unnecessary overrides
Tests performed:
- For all apps affected, test enable/disable button works and submitting
configuration form works: with changes updates message and without changes
'settings unchanged' message.
- avahi
- bind
- cockpit
- SKIP: coquelicot
- datetime
- deluge
- SKIP: diaspora
- ejabberd
- gitweb
- i2p
- infinoted
- ikiwiki
- matrixsynapse
- mediawiki
- minetest
- minidlna
- mldonkey
- mumble
- pagekite
- privoxy
- quassel
- radicale
- roundcube
- SKIP: samba
- searx
- SKIP: shaarli
- shadowsocks
- ssh
- tahoe
- transmission
- FAIL: tt-rss (not installable)
- wireguard
- Deluge test that configuration changes when app is disabled work
- Quassel test that setting the domain works when app is diabled
- Transmission test that setting the domain works when app is diabled
- Ikiwiki create form works properly
- Enable/disable button appears as expected when enabled and when disabled
- Enable/disable button works without Javascript
- Functional tests work for affected apps, Tor and OpenVPN
- AppForm is removed from developer documentation
- Forms reference
- Customizing tutorial
- Test all apps using directory select form
- Transmission
- Deluge
- Visit each template that overrides block configuration and ensure that it is
loaded properly and the display is as expected.
- All apps that use AppView that are not tested above should not have an
enable/disable button. That is JSXC, update, config, firewall, storage, users.
Signed-off-by: Alice Kile <buoyantair@protonmail.com>
Signed-off-by: Sunil Mohan Adapa <sunil@medhas.org>
Reviewed-by: Veiko Aasa <veiko17@disroot.org>
- Use a proper i18n label for 'Forwarders' fields instead of auto-generated one.
- Reuse an already used success message for form submission so that a new
message is not introduced for l10n.
- Minor styling fixes and a yapf run.
Signed-off-by: Sunil Mohan Adapa <sunil@medhas.org>