Argument *text* is an addition in 3.7, while the Calamares 3.3
branch supports Python 3.6 and later. Use the 'backwards compatibility'
name of the parameter, *universal_newlines*.
Cherry-picked from 33961ff6f (in the 3.2 branch, though, Python 3.3
is supported).
- Put this in a method of its own even though it's used
only once, so we can put a good name on it.
- If there are no FS choices (e.g. the sample settings file)
then there is no combo-box, and the check was crashing.
FIXES#2029
- wrangle the test framework so it hands the same data to
two different collections of tests; do KDE neon and FreeBSD
separately so it's clearer which lookups are being done
(and a failure in one doesn't prevent the test of the other).
The distinction CalamaresUtils and Calamares is old-fashioned,
since we can use nested namespaces (and already do) for a lot
of things; make libcalamares/locale/ a bit more consistent
by using namespace Calamares::Locale for everything.
Some distro's let the user change the hostname at will;
others don't, and yet others don't have systemd to change
the hostname with at all. Check if we **can** change the
hostname (as a non-root user), before setting expectations.
- remove "bogus" key when it's not needed
- check for existence of "branding" key in GS before
subscripting it (this happens in tests, where no
GS contents are loaded, but not in regular use,
where startup loads the branding data into GS)
The schema is considered invalid by **some** versions of
configvalidator (or rather, the underlying jsonschema):
```
128: '#definitions/groups' does not match '^[^#]*#?$'
128:
128: Failed validating 'pattern' in metaschema['properties']['definitions']['additionalProperties']['$dynamicRef']['allOf'][0]['properties']['$id']:
128: {'$comment': 'Non-empty fragments not allowed.',
128: '$ref': '#/$defs/uriReferenceString',
128: 'pattern': '^[^#]*#?$'}
```
Remove the `#` from the IDs.
- can't use *bogus* as a key unless that's allowed by
the schema -- and it is not.
- can't supply empty config if there is supposed to be
an object there.
This makes it a little difficult to allow a config-file
that is actually no-configuration-at-all (or only-defaults).
Put in values for *btrfsSwapSubvol* since it isn't a fragile
setting.
Looks like clang-format was applied to this JSON schema
file in 5a14c3c76f, which is
completely borked. Restore from earlier revision, now
with just `machineid` removed.
A stray "machineid" was added, without a type definition,
so the config file -- which doesn't mention machineid at all
-- was considered invalid. Drop it because the key isn't mentioned
anywhere else.
- expand tests to include Serbian variants
- massage detection code to do better filtering based on
country-selection (so "American English" doesn't later
pick "English" with Antigua and Barbados for locale)
starting with 217e4ab4f7
the messages have expanded, no longer fitting in the used ListView model
set the text blocks to expand with text size, add a scrollbar
reduce top spacing so underlying image won't show
set the full requirements listing as default
This looks to be a remnant from moving this to the mount module
which left an undefined variable. The mount options are now part
of the globalstorage and consumed here as mount_options_list.
FIXES#1980
When the checks were done, the widget showing the countdown-timer
and results was not being informed, so it didn't update the
display of the countdown timer or hide the list of problems
(when there are none) so that the welcome image is shown.
FIXES#1974
Adds a new option / configuration keys to `fstab.conf` to
configure how /tmp is created. The example shows how /tmp
is made *tmpfs* on an SSD, or on not-SSD, is just-a-directory.
FIXES#1818
- for QML, this is the easiest way to factor out the message
(no convenient access to free functions, unless I bung the
message into e.g. Branding where I don't think it fits
very well).
Previously, state() just returned a stored state, which changed
via updateState(). However, when updateState() started taking
visibility-of-the-widget into account, it became possible to
de-sync the *apparent* state of the encryption widget, from the
stored one:
- make an encryption widget, which is not visible
- show it.
Now the stored-state takes visibility into account that is
different (hidden, so we end up with a state of Unconfirmed)
from the apparent value (shown and unchecked).
Move the calculation to state() instead, so whenever queried,
it checks the current checks-and-visibility values. Restore
the previously-reverted bit for accepting LUKS partitions.
SEE #1935
SEE #1953
- a hidden checkbox should not be handled the same as a checkbox
that is checked; if there is a setting that can be checked or
unchecked, it can only be checked when visible.
FIXES#1953
- Improve the messages a bit
- Remove useless variable (it is checking that the switch(choice)
statement covers all the cases; let's leave that to the compiler)
esri has several bugs, not showing for all locations, zooming on Americas broken
use osm as default, set preferred, so esri can be fallback
remove Label section for mouse movement, showing coordinates not really useful,
issue with mouse grabbing is back with it
- in 3.3, the legacy values are ignored, so all the old-style
tests behave as if nothing is set at all.
- Some tests used old-style settings, adapt to newer ones.
Some QPainter enum values were deprecated in 5.14, and since we
require 5.15 now, we hit some old code paths with now-deprecated
values; this only shows up when actually enforcing deprecations.
Because of the way Qt interprets the environment variable
LANG, using `sr@latin` or `sr@latn` or `ca@valencia` would
get you `sr` or `ca`, respectively, which isn't an exact
match. Now that Translation has special-handling for
those values of LANG, match with the ID first.
This allows starting Calamares in Serbian (Latin script)
or Catalan (Valencia) for locales that need it. (Qt doesn't
recognize ca@valencia as a variant, since that's a region-
based locale, not country- or script-based)
