FOCAS: porting backlog from the stranded auto/driver-gaps branch (15 items, writes policy decision first, #607) #633
Assignees
dohertj2 (dohertj2)
Clear assignees
dohertj2 (dohertj2)
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: dohertj2/lmxopcua#633
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Porting backlog for the FOCAS driver, from the #607 branch ledger (
docs/plans/2026-10-08-driver-gaps-branch-ledger.md). These issues were auto-closed against PRs merged intoauto/driver-gaps, which never reached master (and predates v3). Each item below was classified STRANDED: no equivalent on master (symbol/file grep plus docs), so the work needs a port-or-drop decision. The branch commit is listed per item (git show <sha>onorigin/auto/driver-gaps); the branch is a design reference, not a cherry-pick, because the driver and the address-space model were rewritten since.Policy decision first: master FOCAS is intentionally READ-ONLY (
WireFocasClient.WriteAsyncreturns BadNotWritable,Wire/WireFocasClient.cs). #268-#271 (writes, wrmacro/wrparam, pmc_wrpmcrng, password/unlock) are safety-critical on real CNCs and need an explicit go/no-go before any port. Everything else needs the managed wire client (Wire/*), not the branch's FWLIB-style code, and CNC hardware orfocas_mockto verify.Source: #607; ledger rows for the original issue numbers below.
Port-or-drop checklist (15 items; effort S/M/L from the ledger):
551494d2). Master has Identity/Axes/Spindle/Program/OperationMode/Timers fixed tree only (docs/drivers/FOCAS.md:213); no ODBST status flags / Port risk: High: master FOCAS is a managed wire client rewrite (Wire/*); branch used FWLIB P/Invoke so must be re-done as wire commands; hardware-only to verify3d9697b9). Timers/CycleSeconds present; parts count absent (partial) / Port risk: High: wire-client rewrite; needs cnc_rdparam/rdtimer decode on hardware3c2c4f29). No modal/override reads / Port risk: High: wire-client rewrite9ec92a90). No tooling/work-offset reads / Port risk: High: wire-client rewritecc757855). No operator message / block text / Port risk: High: wire-client rewrite451b37a6). No DIAG: address area (FocasAddress has PARAM/MACRO/PMC/FIXED only) / Port risk: High: wire rewrite2f3eeecd). No path-number / multi-path support found in FocasWireClient/options / Port risk: High: wire rewrite; hardware63099115). FocasCapabilityMatrix.cs:87 Sixteen_i PMC letters still {X,Y,R,D}; branch adds F,G (4-line change) / Port risk: Trivial, verify against 16i wire4d3ee472). No PMC range coalescing (FocasPmcCoalescer); wire client has pmc_rdpmcrng primitive / Port risk: Medium: fits wire client7f9d6a77). No cnc_rdalmhistry; FocasAlarmProjection is polled cnc_rdalmmsg2 only / Port risk: High: wire rewrite + history decode1bfe8fba). Master FOCAS is intentionally READ-ONLY: WireFocasClient.WriteAsync returns BadNotWritable (Wire/WireFocasClient.cs:13-15,73) / Port risk: High: safety-critical on real CNCs; policy decision needed before any portf48f31cf). wrmacro/wrparam deliberately not exposed in wire client / Port risk: High: as #26854c09d4d). pmc_wrpmcrng not exposed; PMC bit-RMW absent despite stale comment in FocasDriverOptions / Port risk: High: as #26886f3fc27). No password/unlock / Port risk: High: as #268e3d7c65f). No per-part cycle delta / Port risk: Depends on #258Filed 2026-10-08 by the hardware-available issue sweep.
Per-item status (merge
7a9b51aeb626-ablegacy-focas):Left open for the user to decide closure.
Landed in b640 (FOCAS merge
216dc63a+c86204b7): #257 Status/ nodes, #266 PMC read coalescing, #258 parts-count docs; live against focas-sim: new tests pass. Remaining: #259 #260 #261 #263 #264 #267 #272 need wire captures, hardware or design. New: FOCAS native alarms are unroutable -> #647; stale focas-sim fixture image -> #651. Leaving open.Progress from batch b647 (pushed
e7dd7425): #268–#271 recorded as won't-do (FOCAS read-only policy) in docs/drivers/FOCAS.md 'Not ported'; #258 parts count now recommendsMACRO:3901(the Makino refuses PARAM:6711 with EW_FUNC); #259 overrides documented as a read-only PMC recipe (G12 / G14.0-1 / G30); #272 replaced by a virtual-tag recipe; #265 was already on master. Still open: read-only Makino confirmation of #257 / #266 / #258, and the blocked items (modal codes, offsets, operator messages, DIAG:, alarm history, multi-path), which need FOCAS command IDs captured from a non-production control — never probed on the Makino.Read-only live checks on Makino Z-34184 (master
ec375538), FOCAS read calls only, one session, ≥ 2 s polling:Status/flags decode steadily:Auto=3 Run=0 (RESET), no E-stop,Alarm=0(the triage's "permanent alarm=4" is not true today). Open question for someone at the panel:Auto=3is EDIT in the documented table, but the separate operating-mode read reported AUTO throughout.MACRO:3901/3902read 0.0 at Good.pmc_rdpmcrngranges, identical values.255 - vgave 356); corrected to mask the byte, and noted that 128–150 % spindle overrides read negative.Observed, not changed: the block-count read is refused (rc=-7) and re-sent on every program poll (82× in 4 min); the fixed-tree default 250 ms poll is ~28 req/s on this 7-axis control — worth checking the production setting.
Remaining items still need a non-production control (command-id capture).
Closing: every item that can be checked read-only on the Makino is done (see previous comment). The remaining items need command-id capture on a non-production control and stay out of scope under the FOCAS read-only policy; refile individually if such a control becomes available.