#15821
This PR is adding two improvements and fixing two Windows bugs in Fleet
Desktop:
## Improvement
- We are now capturing the stderr of Fleet Desktop. This helped me find
bug (1) below (otherwise the panic output below was hidden from us).
- To reduce complexity I'm removing the "Theme detection" routine
because we made the decision to use the colored icon for both themes...,
see here:
415d1f493b/orbit/cmd/desktop/desktop_windows.go (L21-L27)
## Bug fixes
1. Fleet Desktop icon not showing in the task bar. This was fixed by
updating to use the latest version of `fyne.io/systray`. (See
https://github.com/fyne-io/systray/issues/22#issuecomment-1173157898.)
2. Orbit now properly detects if Fleet Desktop isn't running on Windows.
Bug (1)'s panic output
```
panic: runtime error: invalid memory address or nil pointer dereference
[signal 0xc0000005 code=0x0 addr=0x0 pc=0x72b14b]
goroutine 23 [running]:
fyne.io/systray.(*winTray).setTooltip(0x1eb5d40, {0x126923f?, 0x0?})
/Users/luk/gopath/pkg/mod/fyne.io/systray@v1.10.0/systray_windows.go:260 +0xcb
fyne.io/systray.SetTooltip({0x126923f?, 0x125fc16?})
/Users/luk/gopath/pkg/mod/fyne.io/systray@v1.10.0/systray_windows.go:961 +0x29
main.main.func1()
/Users/luk/fleetdm/git/fleet/orbit/cmd/desktop/desktop.go:103 +0xba
fyne.io/systray.Register.func2()
/Users/luk/gopath/pkg/mod/fyne.io/systray@v1.10.0/systray.go:98 +0x2f
created by fyne.io/systray.Register in goroutine 1
/Users/luk/gopath/pkg/mod/fyne.io/systray@v1.10.0/systray.go:96 +0xb1
```
- [X] Changes file added for user-visible changes in `changes/` or
`orbit/changes/`.
See [Changes
files](https://fleetdm.com/docs/contributing/committing-changes#changes-files)
for more information.
- [X] Manual QA for all new/changed functionality
- For Orbit and Fleet Desktop changes:
- [X] Manual QA must be performed in the three main OSs, macOS, Windows
and Linux.
- [x] Auto-update manual QA, from released version of component to new
version (see [tools/tuf/test](../tools/tuf/test/README.md)).
This should fix#16006.
On Windows when a process is killed by the Task Manager, it is killed
without any signaling, thus the osqueryd processes are left orphaned.
Executing osqueryd (which we do to get host information) was failing
because the lingering processes had a lock on the database file. The
solution implemented in this PR is to kill any pre-existing osqueryd
processes before running osqueryd.
- [X] Changes file added for user-visible changes in `changes/` or
`orbit/changes/`.
- [X] Manual QA for all new/changed functionality
- For Orbit and Fleet Desktop changes:
- [X] Manual QA must be performed in the three main OSs, macOS, Windows
and Linux.
- [X] Auto-update manual QA, from released version of component to new
version (see [tools/tuf/test](../tools/tuf/test/README.md)).
PS: I added a log of the stdout+stderr of osqueryd execution when such
command fails to execute. (This helped me find the root cause.)
```
2024-01-25T11:57:56-08:00 ERR getHostInfo via osquery output= stderr="E0125 11:57:56.744843
7860 shutdown.cpp:79] IO error: Failed to create lock file:
C:\\Program Files\\Orbit\\osquery.db/LOCK: The process cannot access the file because it is
being used by another process.\r\r\n"
```
PPS: I removed some unused exported methods in the `orbit/pkg/platform`
package.
This relates to #10300
# Checklist for submitter
If some of the following don't apply, delete the relevant line.
- [X] Changes file added for user-visible changes in `changes/` or
`orbit/changes/`.
See [Changes
files](https://fleetdm.com/docs/contributing/committing-changes#changes-files)
for more information.
- [X] Manual QA for all new/changed functionality
- For Orbit and Fleet Desktop changes:
- [X] Manual QA must be performed in the three main OSs, macOS, Windows,
and Linux.
- [X] Auto-update manual QA from the released version of the component
to the new version (see [tools/tuf/test](../tools/tuf/test/README.md))
---------
Co-authored-by: Lucas Rodriguez <lucas@fleetdm.com>
* Adding a new synchronization mechanism between fleet-desktop app and Orbit service. Improved windows service teardown to ensure that fleet-desktop does not get force killed without getting signaled. Improved windows process enumeration to avoid unnecessary delays during windows service start and windows service teardown. Updating windows service to reflect service teardown extra time due to synchronization.
* Add safe mkdirall and open
* Use secure as much as possible and merge gomodules for orbit to fleet
* Improve openfile and mkdirall to check for permissiveness instead of equality
* Don't shift
* Fix links
* Address review comments
