v0.6.4

Released: 2026-05-13

Duplicate “Device Reset Detected” banner. Two independent code paths fired the reset banner on a single mid-session reset: the transport’s Reconnected OK event from the helper, AND the parser-side *** Booting detection on the buffer drain that follows the re-attach. The parser path is now suppressed for 5 seconds after a transport-emitted reset, eliminating the duplicate while keeping the parser-side fallback for cases where the helper doesn’t detect the reset (e.g., warm boot with no SWD interruption).
Multi-match RTT control block disambiguation. When the SEGGER RTT magic string appears at more than one address (common: firmware embeds the magic literal in initialized rodata that gets copied to SRAM at boot, alongside the actual control block), LogScope now validates each candidate by parsing it as a CB and applying three checks: (1) parses cleanly with at least one initialized up-buffer, (2) every up-buffer’s pBuffer points into a known RAM range, and (3) every up-buffer’s sName dereferences to a plausible printable ASCII C-string. The candidate that passes wins; if multiple pass or none do, an actionable error surfaces. Mirrors probe-rs’s approach. Reported by issue #23 (custom nRF54L15 board).

“Set Chip Name…” action in connect-failure error toasts. When the helper reports JLinkReadException: Unspecified error (most often caused by connecting with a generic ARM core name instead of the chip-specific J-Link device name), the error toast now offers a one-click action that opens the chip-name picker (with both common-core quick-picks and an “Enter chip name…” free-text option). The same action is also offered on the “No RTT control block found” error (exit code 2) since a wrong device name is one of the common causes there. Eliminates the need to dive into settings.json for the most common direct-memory-path failure. Reported by issue #23.