add basic crash analytics and very light telemetry to dioxus-cli (#4224)

* wip

* wip: basic telemetry

* fix merge

* keep errors for telemetry

* fix upload when queued events are empty

* add per session id and keep track of build timing

* keep track of what commands are run

* capture panics

* keep track of panics

* better path stripping

* better tracing error telemetry

* only sync every week and check settings before sending any telemetry

* keep track of the args passed to commands

* anonymize commands

* more error tolerant

* wip

* wip

* wip

* merge cnflict

* other merg conflict

* migrate everything from telemetry to logging

* update docs

* wip... about to merge trace controller

* lil bit more refactor to clean up tracer

* more performant log writer

* cleanup ci detecting

* wire up to posthog

* re-uploads work

* yoink out the anon from each file

* cleanup sig

* lil bit cleaner

* dont write empty files

* wip log handler

* fmt

* cleanup logging of backtraces

* add proper backtrace support

* use stable reporter

* adhere to posthog format

* fix stacktraces

* use anyhow for capturing backtraces

* I think we're done?

* cleanup logging

* properly collect!

* move out keycode handling

* she works

* typo

* add extra fields

* fix cargo doc

---------

Co-authored-by: Evan Almloff <evanalmloff@gmail.com>

packages/cli/src/devcfg.rs

@@ -8,11 +8,6 @@ pub(crate) fn should_force_entropy() -> bool {
     std::env::var("DIOXUS_FORCE_ENTRY").is_ok()
 }
 
-/// Should the CLI not download any additional tools?
-pub(crate) fn no_downloads() -> bool {
-    std::env::var("NO_DOWNLOADS").is_ok()
-}
-
 /// Should we test the installs?
 #[allow(dead_code)] // -> used in tests only
 pub(crate) fn test_installs() -> bool {