Commit 96f6a949 authored by intrigeri's avatar intrigeri
Browse files

Changelog: import recent changes, in raw format.

parent 5b0ece6a
......@@ -599,6 +599,259 @@ tails (2.0~beta1) unstable; urgency=medium
Synaptic' scenario, in which the APT update can take >5 minutes.
Will-fix: #10403
* Simplify ISO image naming rules.
... by using the base rule we use for Jenkins all the time, except
when building from a tag (i.e. building a release).
Will-fix: #10349
* Use the menu to open Pidgin's account manager
* Retry running whois when "LIMIT EXCEEDED" is in its output.
Will-fix: #10523
* Revert back to using @screen.wait.
We're going to end up restarting Seahorse on segfault. While doing lots
of testing for #9791 I'm fairly confident that I've figured out
Seahorse's failure modes and when it'll segfault. allowing code to be
simplified.
* Make keyid an instance variable in the GnuPG CLI step.
Doing this will allow this variable to be used in the Seahorse
keysyncing steps.
* Start Seahorse with its own function.
We'll also keep track of how Seahorse was opened in case we need to
restart it during the test suite.
* Improve Seahorse key syncing test robustness
* Avoid waiting unneccesarily for the results window if there's an error
* Use more generalized name for menu image's name.
Gedit no longer has a menu bar at the top of its window, causing this
feature to be broken..
* Teminate gpg after 120 seconds have passed.
During my testing I found that it seems that if we're going to get a
successful result it'll happen by the time two minutes have passed.
If the key retrieval neither succeeds nor fails within 2 minutes we'll
terminate the running gnupg process, force a new Tor circuit, and try
fetching again.
* Reduce timeouts in Seahorse to 2 minutes
* explicitly check that there's not an error from Seahorse
* Don't wait unnecessarily.
Since all of the gpg operations will retry after forcing a new Tor
circuit, by the time we reach the step that checks whether the key is in
the keyring the key should be there. We'll gain nothing from waiting two
additional minutes.
* Reduce redundancy with checking for seahorse errors
* Allow step to be retried.
If we don't monitor whether the key has been fetched with the gpg CLI
tool we can move to the next step which used to check whether the key is
in the keyring. By that time it's too late to retry fetching.
Instead let's check with `gpg` for the key. If the key is not in the
keyring within two minutes we'll fail the step, force a new tor circuit,
and try retrieving the key again.
* Remove unneeded return
* s/gsub/sub/
* Pass argument by name
* (Hopefully) more accurate name of function
* More specific variable name
* Typo fix
* Reword scenario title.
* Test suite: move upgrade scenarios to the feature dedicated to them.
* Test suite: fix error message for isohybrid test.
* Update jenkins-tools submodule.
* Stop tweaking /etc/modules.
It's 2015, the kernel should load these things automatically.
Will-Fix: #10609
* Remove obsolete "override our Wheezy-specific package" entries from APT pinning.
They were useful when feature/jessie was an overlay on top of devel,
but it's been not the case for a while.
Will-Fix: #10607
* Install obfs4proxy from deb.torproject.org on Jessie as well.
The package in Jessie is older than the one from deb.tpo, that we currently ship
in Tails 1.x, so let's fetch it from the same place.
Will-Fix: #10605
* Install Tor from deb.torproject.org's stable repo.
The one we currently want (0.2.7) has been released as stable.
Will-Fix: #10604
* Call out to the shell less often in htpdate.service.
On my test system they all are run within the same seconds, but it adds up to
everything else we're doing at the same time.
Will-Fix: #10612
* Clarify status of ttdnsd: we do run it, but via a NetworkManager hook.
Closes: #10608
* Convert config/chroot_local-hooks/21-gdm_unit_file to a drop-in override.
I see no reason to patch it at build time the way we did it.
Will-Fix: #10606
* Wrap Totem with torsocks when it's run via its D-Bus service too.
Will-Fix: #10603
* 09-torsocks-apps: make regexps a bit stricter.
* Have systemd hardening let Tor modify its configuration.
Tor Launcher needs that.
Closes: #10696
* Tag `@doc` scenarios that rely on the documentation.
Will-fix: #10706
* Restore image into broken symlink.
The symlink's target image was removed.
* Remove test of the removed restricted network detector.
* Remove silly executable bit.
* s/syslog/journal/
* Refactor the PAUSE_ON_FAIL functionality.
A well-placed `pause()` can act as a breakpoint when debugging.
* Kill + inline strangely named function.
And fix the code so we get the proper middle, i.e. measure from the
top's bottom to the bottom's top (hah!).
* Remove old workaround that seems obsolete.
... at least on my system, that previously could reproduce it.
* Adapt test to how we disable networking in feature/jessie.
See `stop_and_disable_NM()` in the `tails-mac-spoof` script.
* Bump some images for feature/jessie.
* Kill Vidalia before we restart Tor.
Otherwise Vidalia will be running and showing errors while we make
sure that Tor bootstraps, which could take a while.
* Remove unnecessary waiting code.
Earlier we already ensure that Tor has bootstrapped, which implies
that the control port must be up.
* Remove obsolete comment.
The referred-to comment in checks.feature is not there any more, and
in this feature we use snapshots any way (although not as freely as
we'd like).
* Use snapshot.
This scenario is not using a filesystem share, so unlike some other
scenarios in this feature there's nothing preventing us.
* Remove superfluous part of test.
We still do the essential part i.e. test that Totem can open videos
over Tor. The removed part is essentially a test of the GUI, and what
a confusing and hard-to-discover GUI it is! I doubt users will ever
find out how to do it, so I fail to see why we should test it (and
waste time dealing with the robustness issue that plague this test).
* Add retry magic to the Totem HTTPS scenario.
* Use journalctl instead of (non-existing) syslog.
* Adapt test for new Tails Installer.
The USBCreateLiveUSB.png image cannot be relied upon for waiting on
the main installer window, since it can be greyed out. Also, the error
message when a device is not upgradable appears without even clicking
the install button.
* Remove comment about OCR, and commented OCR code.
Sikuli's OCR capabilities are way too bad, so let's forget about this.
* Introduce a new XML-based iptables parser.
The old `iptables_parse()` way relies on regex parsing the output of
the `iptables` output, which of course is not very reliable. In Jessie
the `ip6tables` output has changed and differs (if there are no
options the "opt" column will be empty, and not have a "--"). The new
parser fixes this.
The goal will be to convert all `iptables_parse()` instances to the
new parser.
Will-fix: #9704
* Convert `iptables_parse()` instance.
* Convert `iptables_parse()` instance.
* Convert `iptables_parse()` instance.
* Convert `iptables_parse()` instance.
* Kill the now unused `iptables_parse()`.
* Lower waiting time for USB installation in the test suite.
So far we were waiting up to one hour, which is just the same as our
Jenkins inactivity timeout, so in practice when Tails Installer fails
and displays an error message, instead of reporting that the job failed
(which is the point of the exercise) we abort the job due to this
timeout (which communicates waaaay less clearly to me, with my Tails
Installer developer hat, that there's probably a bug I should fix, as
opposed to "here's a transient failure caused by instabilities of our
Jenkins setup"). My understanding is that this is what prevented the
Tails Installer bug described on #10717 from being identified earlier,
and I don't want to see that happen again.
IIRC this 1h timeout was relevant for running with nested virtualization
on slightly old (ThinkPad X200) hardware. I don't think anyone does that
anymore, and I would bet that on such hardware, lots of other bits of
our test suite will time out anyway.
Fix-committed: #10718
* Mark apt-get scenario as fragile.
Refs: #10496
* Flag online scenarios as fragile.
They all rely on the `wait_until_tor_is_working` helper which has lately
proven to be still buggy. That's tracked by ticket #10497.
* Flag a online scenario in Jessie as fragile.
It depends on the `wait_until_tor_is_working` helper which is buggy.
Refs: #10497
* Mark as fragile all tests that rely on Tails Installer.
Refs: #10720
* Test suite: fix steps that shutdown/reboot via the applet.
When we hide the cursor, they often fail: the menu appears and
disappears immediately, before the shutdown or reboot button
is located.
* Use `systemctl mask' instead of manually symlinking, and thus avoid unneeded daemon-reload.
`systemctl daemon-reload' is a very big hammer: while it's running, socket
activation, D-Bus activation, and more systemd functionality are disabled.
In this case, as long as we use only systemctl {stop, disable, mask}, then we
don't need to do a global daemon reload, so let's not take the risk.
* Fix typo.
* Relax permissions on Tor and its child processes, to fix obfs4proxy startup.
Closes: #10724
* Install apparmor-profiles-extra from jessie-backports, that grants Totem access to DVD readers (refs: #10455), and fixes a freeze when trying to disable visual effects (refs: #9990).
* Fix shell syntax in hook that detects leftover .orig files.
* Unfuzzy apparmor-adjust-pidgin-profile.diff.
* Unfuzzy chroot_local-patches/unmute_alsa_channels.patch.
* Fix shell syntax error.
In commit:3e85f97 I've fixed one such error and introduced another one
("git add -p" :)
* Update a test suite image for Jessie.
* Fix scenario name.
This one does not test ~/.gnupg/.
* Test suite: don't reboot when not needed.
These lines somehow survived or were re-introduced via merge
conflict resolution. That's the only explanation I have for them.
* Optimize memory erasure anti-test: block the boot to save CPU on the host.
* Add the bugfix-9713-electrum-2.5 APT overlay.
Will-fix: #9713
* Fix pinnings for electrum and its dependencies.
Now we install a custom Wheezy backport of electrum from our own
repo. Note that python-pbkdf2 is not even in Wheezy Backports, so we
install it from Debian Jessie.
* Make config appropriate for electrum 2.5.x.
For now, let's also drop "auto connect" and server until we've decided
on a server we can trust.
* Flag forgoten online scenarios as fragile.
Refs: #10497
* Refresh patch against Tor's AppArmor profile to apply on top of 0.2.7.6-1.
* Upgrade Tor Browser to 5.0.5.
Will-fix: #10751
* Warn if the claws-mail persistence is enabled when starting icedove.
Refs: #10458
* Warn if claws-mail persistence is activated AND it contains a accountrc file.
Otherwise people will get the notification every time they start Icedove
during the 1.8 lifetime.
Will-fix: #10458
* Update AdBlock Plus patterns.
* Add Icedove to GNOME launcher, remove Claws-Mail
* Empty the list of APT overlays: they were merged
* Update PO files.
* Install Electrum from jessie-backports.
That is, currently: 2.5.4-1~bpo8+1. This is older than what we'll
be shipping in Tails 1.8 (#9713), and we'll need a patch that's
in 2.5.4-2, but for now it's still better than Jessie's 1.9.8-4,
presumably.
refs: #10754
* Update changelog for 1.8.
* Updating SquashFS sort file
* Fix ticket number
* Add security updates to the 1.8 changelog entry.
* Test suite: drop unused steps.
We're not doing anything with that file, no need to copy it.
* Set base branch to devel.
* Update POT and PO files.
-- Tails developers <tails@boum.org> Thu, 19 Nov 2015 16:01:19 +0000
tails (1.8) unstable; urgency=medium
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment