A recent branch carefully called Prober.get_storage() whenever a udev
block device event was seen. But Prober.get_storage() cached the results
from probert rendering that useless. Rip the caching and some other
complexity out of the prober class.
In dry-run mode, treat SUBIQUITY_DEBUG as a comma-separated list of
things to fail:
* install-fail: curtin install fails
* copy-logs-fail: the postinstall copying of logs fails
* bpfail-full: full block device probing fails
* bpfail-restricted: restricted block device probing fails
In addition, control-u makes the UI crash in dry-run mode.
(This is all for testing of crash report generation but does not depend
on that).
Log to subiquity-debug.log.$PID instead of just subiquity-debug.log.
When we implement error reports we'll obviously attach the log to the
error report. There's no point adding log messages from previous runs of
subiquity.
This will be used for the drop to shell functionality and also to allow
viewing an error report before we submit it to daisy.
It is almost but not quite surprisingly easy (I added a long-ish comment
about the difficulties).
For whatever reason, urwid puts the terminal into cbreak mode during
initialization. If we put the terminal into raw mode instead, then we
don't have to ignore SIGINT and SIGQUIT, which is good, because when we
support dropping to a subshell we don't want to run that subshell with
those signals ignored, because that is extremely confusing.
This also lets me dump the code that puts the terminal into raw mode
during keyboard detection.
Set fallback=True on gettext.translation() call such that it returns
NullTranslation object. This means subiquity/console-conf can start,
even if .mo files are not build in the tree, or are outright missing.
it seems netlink routing table messages are more of a notifications of
changes being explicitly made vs notifications of all changes to
routing.
for https://bugs.launchpad.net/subiquity/+bug/1837822
When allocating space for a cell colspan > 1, preferentially
allocate space to columns that have no colspan = 1 cells. If you
have content like this (content padded to 10 and 20 chars to make
calculations easier):
[[(2, "longer content567890"), "shorter890"],
["shorter890", (2, "longer content567890")]]
Taking only colspan = 1 cells into account, the widths of the
columns are (10, 0, 10).
To fit the first longer content in, the current algorithm expands
the first two columns equally, so the widths become (15, 5, 10).
Then to fit the second longer content in, the second two columns
are widened equally (or nearly so), resulting in final widths
of (15, 8, 12) and a table layout like this:
+-------------------------+--------------+
| longer content567890 | shorter890 |
+-----------------+-------+--------------+
| shorter890 | longer content567890 |
+-----------------+----------------------+
which is bizarrely asymmetrical given the input and wider than it
needs to be.
If instead we preferentially widen columns that have no width
from colspan = 1 columns, we get this layout:
+----------------------+------------+
| longer content567890 | shorter890 |
+------------+---------+------------+
| shorter890 | longer content567890 |
+------------+----------------------+
Which is much more in line with my expectations.
So I can write a unittest more easily.
This involves shuffling around how locale changes are done but as my new
design document says the "controller also manages the relationship
between the outside world and the model and views" this does make things
more consistent.
Normally netplan.Config().config is a dict as produced by
yaml.safe_load(). However, on systems without any netplan configs that
is not true. The subiquity code that calls into this function expects
.config to be a dictionary. When it's not
subiquitycore/models/network.py produces traceback that "list does not
have get method" when trying to get 'network' key.
my best guess at why CI is currently hanging sometimes is that the
'next-screen' signal is being sent too often / too early. add some
logging around this so that we will be able to confirm / deny this by
reading the logs.
Mark all devices with DHCP enabled as pending when the config is
applied, and mark them as timed out ten seconds later if they have not
received any addresses.
The visible impact of this is that the addresses for a NIC are now
displayed underneath its name and a long "info line" (i.e. mac / model /
vendor) for a nic no longer distorts its table.
This lays the groundwork for better handling of NICs that fail to dhcp
and things like that.