vehicle_vwegolf — proposing a decode/TX source split to cut cross-fork merge conflicts
Two independent forks are now actively working the vwegolf (VW e-Golf) component, mine (climate/BAP, charging next) and SCjona's (the ignition-status / v.e.on / v.e.awake work that just merged). We're both editing the same file, vehicle_vwegolf.cpp, and it's starting to hurt. Every time I sync master I eat a batch of moved-function-body conflicts, 11 on the last big sync, all mechanical, all resolved the same way. That cost only grows as the file grows and as two forks diverge on the same lines. In my branch I've already split the component along the obvious seam: decode/RX stays in vehicle_vwegolf.cpp, and the whole KCAN TX/actuation path, OCU heartbeat, NM wake, the BAP clima bursts, the Ticker1 sub-tasks, moved into a new vwegolf_tx.cpp, about 500 lines out of the monolith. No behavior change in that commit, it's a pure move: https://github.com/slothish/ovms3-VWeGolf/commit/b5a17046dfc1d5f223fd97b1a64... I'd like to land that split upstream on its own, before the file grows further — charging control is next and it lands entirely on the TX side. The win is shared, not just mine: anyone touching the component stops colliding with everyone else's decode edits, and future foreign PRs rebase onto a layout that keeps RX and TX apart. If there's appetite I'll submit it as a standalone no-behavior-change PR off current master, with the concurrency fixes I stacked on top kept as separate small PRs so the move stays clean to review. Happy to coordinate timing with SCjona so it lands in a quiet window and we both rebase once. For transparency, like the rest of my e-Golf work this is heavily AI-assisted and leans on the native tests, compiler feedback, and real CAN captures for validation — the split itself is verified behavior-preserving by the component's test suite.
Andreas & SCjona,
you are free to re-/organise your vehicle module source as you like.
Source restructuring (and indentation/bracket style unifications) shall always be separate PRs without functional changes, as all functional changes need to be revertible individually.
Regards, Michael
Am 23.07.26 um 23:51 schrieb Andreas Jansson:
Two independent forks are now actively working the vwegolf (VW e-Golf) component, mine (climate/BAP, charging next) and SCjona's (the ignition-status / v.e.on / v.e.awake work that just merged). We're both editing the same file, vehicle_vwegolf.cpp, and it's starting to hurt.
Every time I sync master I eat a batch of moved-function-body conflicts, 11 on the last big sync, all mechanical, all resolved the same way. That cost only grows as the file grows and as two forks diverge on the same lines. In my branch I've already split the component along the obvious seam: decode/RX stays in vehicle_vwegolf.cpp, and the whole KCAN TX/actuation path, OCU heartbeat, NM wake, the BAP clima bursts, the Ticker1 sub-tasks, moved into a new vwegolf_tx.cpp, about 500 lines out of the monolith. No behavior change in that commit, it's a pure move:
https://github.com/slothish/ovms3-VWeGolf/commit/b5a17046dfc1d5f223fd97b1a64...
I'd like to land that split upstream on its own, before the file grows further — charging control is next and it lands entirely on the TX side. The win is shared, not just mine: anyone touching the component stops colliding with everyone else's decode edits, and future foreign PRs rebase onto a layout that keeps RX and TX apart.
If there's appetite I'll submit it as a standalone no-behavior-change PR off current master, with the concurrency fixes I stacked on top kept as separate small PRs so the move stays clean to review. Happy to coordinate timing with SCjona so it lands in a quiet window and we both rebase once.
For transparency, like the rest of my e-Golf work this is heavily AI-assisted and leans on the native tests, compiler feedback, and real CAN captures for validation — the split itself is verified behavior-preserving by the component's test suite.
OvmsDev mailing list -- ovmsdev@lists.openvehicles.com To unsubscribe send an email to ovmsdev-leave@lists.openvehicles.com
-- Michael Balzer * Am Rahmen 5 * D-58313 Herdecke Fon 02330 9104094 * Handy 0176 20698926
participants (2)
-
Andreas Jansson -
Michael Balzer