Add support for the Wingtech CT2MHS01 hotspot, a Qualcomm mdm9650-based
device with a screen available for US$15-35. This device is often used
as a base platform for while labeled versions like the T-Mobile TMOHS1.
AT&T branded versions of the hotspot seem to be the most abundant.
The device has a framebuffer-driven screen at /dev/fb0 that behaves
similarly to the Orbic RC400L, although the userspace program
`displaygui` refreshes the screen significantly more often than on the
Orbic. This causes the green line on the screen to subtly flicker and
only be displayed during some frames. Subsequent work to fully control
the display without removing the OEM interface is desired.
A lot of the bug report we receive are about the web UI or the installer
failing, and there things like capture date just don't matter. We could
create separate templates for these types of bugs, but I'd think it's
probably better to just have one textbox with a few "reminder" questions
that are all optional.
Feature request template I think doesn't have this issue.
Also allow the creation of blank issues, because some issues are more
related to CI or devenv and don't neatly fit in any category. Let's just
hope nobody abuses that?
This commit introduces release automation triggered by button clicks in
Github Actions, guarded by a check on whether all the Cargo.toml files
contain the same version string.
On PRs, changes to documentation no longer trigger code tests.
Similarly, changes to code that don't update documentation do not
trigger documentation tests. Changes that fail at the `cargo check`
stage abort early to prevent lengthy CI builds of the installer and
firmware.
Commits on the `main` branch always run the full test suite regardless
of what changed.
Releases also run the full check, test, build and publish suite.
Due to a refactor in https://github.com/EFForg/rayhunter/pull/350, we
had to move more into the shared codepath of StopRecording. The todo!()
used to be unreachable when it was just in the stop-recording endpoint.
Fix#367
