Happy to pick up the remaining piece of this — pure-decimal node IDs at the CLI.

From the thread so far, the consensus rule would be:

!deadbeef / 0xdeadbeef → hex node id (already works since 464: allow for 0x node prefix values #731)
digits-only, no prefix → decimal node number
anything else → unchanged behaviour

While implementing this I noticed two things about the current hex branch in _sendPacket() (isinstance(destinationId, str) and len(destinationId) >= 8 → int(destinationId[-8:], 16)):

digits-only input like --dest 12345678 currently parses as 0x12345678 (=305419896) — under the rule above it would become decimal instead, so I'd flag that as an intentional change in the PR/docs.
any non-hex string of 8+ chars (e.g. a long node name) raises a raw ValueError before the NodeDB lookup can run — my branch turns that into the same friendly "not found in DB" exit the rest of the path uses.

On juergenRe's ambiguity point: I checked and destination strings are never resolved against node names today (self.nodes is keyed by !hex user ids), so there's no name-collision surface. Digit strings shorter than 8 characters currently just error out ("not found in DB"), so decimal support adds a capability; longer ones are the intentional behaviour change flagged in point 1.

I have a working branch with tests covering all input classes (prefixed/bare hex, decimals incl. range limits, unicode digit lookalikes, overlong inputs) verified against a real node, and will open a PR here as soon as you're happy with the direction.

Accept node.id in decimal and hex #464

Description

Activity

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions