This change adjusts the layout for the progress bar, write button and
cancellation buttons such that they are all placed on the same layout
'row', with the progress bar stacked as before, and the
write/cancel/cancel buttons stacked as a column.
This is not ideal, but probably as sensible as we can get inside this
layout paradigm.
1) Resize the window back to 680dip default widths. We don't need the
extra space now.
2) Reduce Row spacing within the grid layout. We need all the space we
can get.
3) Assign layouts to explicit cells, at least for selection options and
write. This layout isn't really scaling to the amount of data we want
to provide, but we'll make do for now.
4) Mark selection buttons as accessibility ignored when the hwpopup is
active.
Introduce 4 filtering types per device spec:
- "exclusive", which only includes OS' that match one of the specified
tags, and discards any OS that is untagged.
- "exclusive_prefix", which only includes OS' that prefix-match one of
the specified tags, discarding any OS that is untagged.
- "inclusive", which discards any OS that doesn't include one of the
specified tags, but includes untagged images.
- "inclusive_prefix", which discards any OS that doesn't prefix match
one of the specified tags, but includes untagged images.
Removed preferred widths for OS, device buttons, and set one for the
storage button. This allows the storage label (generally the longest)
to fit on the control face reliably.
Prior to this change, you could enter a submenu of the OS selector,
close the window, and then change device filter - presenting stale information
when you went back in to the OS selector.
Work around this by resetting the OS selector sequence when you close
the OS selector window.
Rather than a drop down dialog, which could present users with images
that may not run on their hardware, allow selection of Raspberry Pi as a
first stage. If users adopt this feature, they are presented with a
subset of images that we know will actually run on their hardware.
This is achieved by leveraging @maxnet's excellent OS filtering scheme.
Future work will attach image and description support to this OS list.
Add " %F" to the Exec parameter in the rpi-imager.desktop file so that
Linux desktop environments know that rpi-imager can be used to open
files. After this change, Imager appears within the list of "Open With"
applications in GNOME, and you can set it so double-clicking image files
always opens rpi-imager.
If you load a custom image by running `rpi-imager my_image.img.xz` then
the Advanced Options button is not shown. This is triggered if you set
a file association to load rpi-imager and then double-click an image
file.
The fix is simply to set the visibility on this button on start-up
(custom images set on the command-line are loaded before this).
Fixes#451
If set(ZLIB_INCLUDE_PATH value) is called without setting cache,
and find_package(zlib) is called twice, the second call fails.
Behavior described here: https://gitlab.kitware.com/cmake/cmake/-/issues/21343
It seems that the new libcurl version also calls find_package(zlib)
so the second call by libarchive fails. And without zlib support
handling .zip files that use deflate compression does not work.
Only affected Windows, because we are only using a bundled zlib
there.
On Linux system libarchive and system zlib is used.
On Mac OS X bundled libarchive, but system zlib is used.
Ref #586
Appearently wlanapi.dll (which we use to borrow the wlan password
from the computer running Imager) is not always present on the
server editions of Windows, so only load it if it exists.
Ref #584
Otherwise it may still be doing it during ejection, resulting
in read IO errors in dmesg, which are harmless by themselves
but may confuse users.
Also tell udisks2 not to analyze disk by setting 'HintAuto'
property on BlockDevice object to 'false'
