We fix store autoselection, which selects an empty value when one
drive is selected and then ejected, leaving one drive that is
supposed to be autoselectable. Now it instead properly autoselects the
last drive.
Change-Type: patch
Changelog-Entry: Stop store autoselection from selecting empty values on
drive ejection.
We move the memoize function to `lib/shared/utils.js` and expose it to
modules across the project.
Change-Type: patch
Changelog-Entry: Move memoize function to shared utils.
We add the missing `name` parameter to the `verifyNoNilFields` JSDoc
example.
Change-Type: patch
Changelog-Entry: Add missing name param to verifyNoNilFields JSDoc example.
We lay the foundation for multi-selecting drives by implementing it into
the `store` and relevant modules interacting with the `store`.
Change-Type: patch
Changelog-Entry: Add drive multi-selection to the store.
We refactor and simplify the selection-state `.clear()` by removing the
options argument.
Change-Type: patch
Changelog-Entry: Remove `selectionState.clear()` options argument to
simplify.
We make the nil-checking of store state fields generic through a
`verifyNoNilFields` function that throws an error if any fields are
nil.
Change-Type: patch
Changelog-Entry: Consolidate store state nil-checking with helper
function.
We warn the user when they select a large drive to confirm they want to
flash in case the device is important.
Fixes: https://github.com/resin-io/etcher/issues/1916
Change-Type: patch
Changelog-Entry: Warn the user on selection of large drives.
We remove a piece of code checking whether `_.keys` returns any non-string
values in its array, but per the Lodash documentation `_.keys` always returns an
array of strings.
Change-Type: patch
Changelog-Entry: Remove stale `invalidKey` check in store.
This removes the remnants of the unused "robot protocol",
which was previously used by the CLI, but has been superseded by IPC.
Change-Type: minor
Changelog-Entry: Remove unused robot protocol
We upgrade eslint for better support of the object-curly-newline
options, and change the quantity of elements required before the linter
enforces newlines on objects, in addition we apply `consistent` to avoid
errors when there are newlines in objects with less than 2 elements.
```
const { field } = object
const {
field
} = object
const {
field1,
field2
} = object
````
Change-Type: patch
Changelog-Entry: Upgrade eslint to use object-curly-newline options.
We add an icon next to the drive size that is displayed when there is a
drive-image compatibility status message available. We display the first
one in the list and importance is then enforced by the order they are
added to the list in `drive-constraints`.
Change-Type: patch
Changelog-Entry: Add icon next to drive size when compatibility warnings exist.
We replace the lodash templates with arrow-functions and change the
single-argument object into multiple arguments.
Fixes#1810Closes#2006
Change-Type: patch
Changelog-Entry: Replace Lodash templates with arrow-functions.
Due to some Windows systems missing certain C runtime libraries
(Visual C/C++ 2012 / 2015 Redistributables), we ignore errors when loading
this module until we can ensure distribution of those along with it.
Change-Type: patch
Changelog-Entry: Fix "The specified module could not be found" on Windows
Fixes#1956
The Makefile current has logic to disable updates when building deb or
rpm packages. To make the Concourse pipeline transition easier, the
logic that disables updates on deb and rpm has been moved to the main
application code.
Change-Type: patch
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
Etcher currently elevates a child writer proxy that itself spawns the
Etcher CLI in robot mode, parses the output, and proxies those messages
to the GUI application over IPC.
After these set of changes, Etcher elevates a single child writer
process that directly communicates back with the GUI using IPC. The main
purpose behind these changes is to simplify the overall architecture and
fix various issues caused by the current complex child process tree.
Here's a summary of the changes:
- Stop wrapping the Etcher CLI to perform writing
- Remove the robot option from the Etcher CLI (along with related
documentation)
- Elevate a new `child-write.js` standalone executable
- Move the relevant bits of `lib/child-writer` to the `image-writer` GUI
module
- Remove the `lib/child-writer` directory
- Add a new "Child died unexpectedly" Mixpanel event
- Floor state percentage in the flash state model
The above changes made is possible to tackle all the remaining issues
where the writer process would remain alive even if the parent died.
Change-Type: patch
Changelog-Entry: Ensure the writer process dies when the GUI application is killed.
See: https://github.com/resin-io/etcher/pull/1873
See: https://github.com/resin-io/etcher/pull/1843
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
This is a strange one. On Windows, putting a space before the
double-ampersand command concatenator makes the environment variable
value contain a trailing space. So for something like `set foo=bar &&
...` the variable `foo` will be `'bar '` instead of `'bar'`.
Change-Type: patch
Changelog-Entry: Fix trailing space in environment variables during Windows elevation.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
By far WoeUSB is one of the few applications that support Windows image
and runs on GNU/Linux so I assume it is worth to mention it. Additional
line wrapping is made to comply to the code conventions.
Signed-off-by: 林博仁 <Buo.Ren.Lin@gmail.com>
This is the first step towards full usbboot Windows support. The driver
selector dialog will now display disabled devices to represent Compute
Modules even when Windows drivers are not installed to act on them.
These drives will state "Missing drivers."
Change-Type: minor
Changelog-Entry: Display connected Compute Modules even if Windows doesn't have the necessary drivers to act on them.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
Various GNU/Linux distributions require root access to be able to open
USB devices. This means that Etcher would need to be ran as root, which
is now not possible on Wayland based systems.
We declare usbboot as unsupported on GNU/Linux from the time being, but
we are currently working on a solution.
Change-Type: patch
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
This implements an SDK.Scanner which handles any given
adapters and manages the scans. This change enables continuous
scanning without the need to `.scan()` scheduling in other places.
Change-Type: minor
We fix the DEBUG environment variable by setting it on the `electron.remote`
instead, and we also move the code to `lib/gui/app.js` and away from
`lib/gui/index.html`.
Changelog-Entry: Set the DEBUG environment variable on the remote
electron process.
Change-Type: patch
* use debug.enabled, fix drivelist env setting
* fix: set debug env variable on remote electron process
This is a workaround to prevent the USB device from disappearing after
the file server phase, until the resulting block device comes up.
By adding a delay after the file server phase, we prevent the USB
scanner from getting triggered again, therefore keeping the current USB
device visible in the drive selector modal.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
This commit re-architects the usbboot adapter to prepare the drives in
the background, while emitting scan results every 2s, where each drive
has a `progress` percentage property.
Change-Type: minor
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
We prompt the user with a sudo-prompt upon launch on Linux-based systems
to ensure the program has enough permissions for features needed
throughout the program's runtime.
Changelog-Entry: Add a sudo-prompt upon launch on Linux-based systems.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
* refactor(SDK): make adaptor scan functions event based
This change will allow us to start emitting progress events out of the
adaptors scan functions.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
* refactor(sdk): Make adapters event emitters
The combination of bus number and device address is the only way to
uniquely identify a USB device, so we'll use that for the `device` and
`raw` properties.
Also, we store the USB vendor and product IDs as properties of the
drives, since they will be handy when implementing the prepare function.
Change-Type: patch
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
Consider the following scenario:
- Usbboot runs successfully on a device
- Before the block device gets a chance to appear, we run usbboot again
If we're fast enough, usbboot will try to claim the device interface,
but then the drive might not be there anymore, causing a
`LIBUSB_ERROR_NO_DEVICE`.
This commit addresses that scenario, and simply ignores the drive.
Change-Type: patch
Changelog-Entry: Fix `LIBUSB_ERROR_NO_DEVICE` error at the end of usbboot.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
We currently ship with `bootcode.bin` and `start.elf` from the Raspberry
Pi Foundation, which provide a writing speed of about 6 MB/s. This PR
includes new boot files by resin.io that boost the speed to ~20 MB/s.
Change-Type: minor
Changelog-Entry: Increase the flashing speed of usbboot discovered devices.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
This commit handles errors that can come up when unplugging the drive
halfway through the process. After tons of experimentation, the errors
than seem to occur are:
- `LIBUSB_TRANSFER_CANCELLED`
- `LIBUSB_ERROR_NO_DEVICE`
When these errors happen, we can omit the drive, and also not try to
close it, since given the device is no longer there, the close operation
bails out with a strange error message.
Change-Type: patch
Changelog-Entry: Gracefully handle scenarios where a USB drive is disconnected halfway through the usbboot procedure.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
We have a debug message that prints `device.name`, which is not a valid
property, and therefore the debug logs show `undefined` instead of the
USB id pair.
Change-Type: patch
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
Some devices don't have a size, like USB devices in the usbboot adaptor.
The `.isDriveLargeEnough()` correctly returns `false` in this case,
however we don't want to show the `TOO SMALL` badge for aesthetics
purposes.
So if a drive has a size that equals `null`, we don't allow such drive
to be selected, and we don't show a badge for it.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
This commit introduces a boolean `disabled` property rather than a
`pending` flag. Making this distinction clearer means that we can now
treat pending drives in different ways needed to improve the usbboot
experience.
Also, for usbboot, this commit removes the "pending" badge and uses a
more descriptive drive description instead.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
- Add a nice icon in the drive selector dialog when a device has been
discovered through usbboot
- Change the name of usbboot-initialized devices to "Compute Module"
Change-Type: minor
Changelog-Entry: Add eye candy to usbboot initialized devices.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
We experienced timeouts when sending big files (ie ~14 MBs). Setting the
timeout to 0 makes the timeout infinite.
Change-Type: patch
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
