Root is always selected, can't be unselected, and has its own explicit
constructor and name. This resolves issue reported where unchecking
all *visible* groups caused the root to be unchecked, after which
hidden-but-still-selected subgroups were not installed.
Reported by crazy@
If a subgroup is hidden, then it should be considered
selected if its parent is selected or partially-selected.
If the parent group is totally unselected, then the hidden
subgroup shouldn't be installed either. This allows putting
required-packages into a group, without cluttering the
interface.
FIXES#864
While walking up the tree, only switch the selectedness states
of parents with children. This avoids the case where a parent
has a first subgroup that is hidden -- in which case the
parent ends up with no children, and is unselected even though
it is marked as selected in the config file.
FIXES#864
- m_groups is only set to a non-nullptr value when data is received
and fully processed,
- avoid nullptr dereference when paging *back* from a netinstall
page that hasn't loaded groups data.
FIXES#859
- Document netinstall.conf a little,
- Add setting *required* which influences whether next is enabled or not
in case of missing or corrupt data,
- Enable *next* button only once some (any!) data is received.
This can be used to disallow stepping past the netinstall step when
there is no data (e.g. internet has failed between the welcome page
and the netinstall page).
- If there's no scripts involved in a package for netinstall, just
name it without the scripts; this lets the packages module
optimize to fewer package manager calls.
- Each element of the (list) packageOperations needs to be a
package operation, which is a dictionary with keys (identifying actions)
and lists of packages (which may be strings or script-info dictionaries).
In the timezone widget, locations in the far north -- Inuvik,
Thule, Longyearbyen -- were displayed too far south, because the
map location calculation assumes a linear gradient, which places
90 degrees north at about 70 degrees. Change calculation to
pretend the world is flat south of 62 degrees north, and then
'bend' the remaining 28 degrees of latitude to the top of the
image. This puts most places in the right spot, although Yellowknife
is now on the south shore of Great Slave.
Fort Nelson should be north of Dawson Creek, too -- the math still
needs a little work.
While here, put Antarctica in the south, otherwise Rothera keeps
showing up in Greenland.
- Introduce a map 'passwordRequirements' in users.conf,
which is a list of named requirements. There are only
two settings right now, min and max length, but
additional checks can easily be added in UsersPage.cpp
by defining additional lambda's to check the given
password string.
- Add PasswordCheck instances as needed, with functions
to check acceptability and to produce messages on rejection.
- Documentation in the users.conf file itself.
- In passing, refactor setting of pixmaps on labels.
FIXES#790
- This is only a partial solution to warnings caused by third-party
code, since #including the headers from other sources won't apply
the warning-suppressions.
- Flags are not applied when building the source as part of a larger
target, but are on re-building just one object (it seems -- CMake
issue to track down).
Thanks to Kevin Kofler for pointing out what I'd forgotten about
source-file flags. While at it, introduce a generic mechanism for
suppressing warnings in third-party code.
Mostly reverts 4930484931
- include full license headers,
- copied from
- repo: https://github.com/snowwlex/QtWaitingSpinner
- rev: bb8f8987ca19406dc75704eb382ab52e981b773f
This revision *does not build* because the files have been renamed.
- Introduce a map 'passwordRequirements' in users.conf,
which is a list of named requirements. There are only
two settings right now, min and max length, but
additional checks can easily be added in UsersPage.cpp
by defining additional lambda's to check the given
password string.
- Add PasswordCheck instances as needed, with functions
to check acceptability and to produce messages on rejection.
- Documentation in the users.conf file itself.
- In passing, refactor setting of pixmaps on labels.
FIXES#790
There is no need to force folks haing a random group
bc that is a Distro think. SDDM/GDM works just fine without
having a group for the user to autologin.
Just setup a group in users.conf .. is why we have a configuration
option for that.
Resolves issue where 'linuxswap' is translated to 'Linux-Swap',
for instance. FileSystem::name() provides a translated name,
not an untranslated one.
This should move to KPMCore.
FIXES#797
Resolves issue where 'linuxswap' is translated to 'Linux-Swap',
for instance. FileSystem::name() provides a translated name,
not an untranslated one.
This should move to KPMCore.
FIXES#797
There is no need to force folks haing a random group
bc that is a Distro think. SDDM/GDM works just fine without
having a group for the user to autologin.
Just setup a group in users.conf .. is why we have a configuration
option for that.
- deleteLater() doesn't like nullptr (produces a warning, but is harmless)
- reparenting across threads doesn't work, comment on that but leave
it in, since this may be relevant for memory management.
- deleteLater() doesn't like nullptr (produces a warning, but is harmless)
- reparenting across threads doesn't work, comment on that but leave
it in, since this may be relevant for memory management.
- both NetInstall (group data) and Locale (GeoIP) use network
data returned as a source of YAML data. Try to explain
parsing errors for both.
FIXES#786
When NetInstall receives YAML data, handle parser errors more gracefully:
show line and column, but because it's network data (not in a local file),
do some work to print out the actual data received.
FIXES#786
- both NetInstall (group data) and Locale (GeoIP) use network
data returned as a source of YAML data. Try to explain
parsing errors for both.
FIXES#786
When NetInstall receives YAML data, handle parser errors more gracefully:
show line and column, but because it's network data (not in a local file),
do some work to print out the actual data received.
FIXES#786
