roa2web-service-auto

romfast/roa2web-service-auto

Fork 0

Commit Graph

Author	SHA1	Message	Date
Claude Agent	6aceb85bf1	docs(service-auto): design review + VFP audit (Correction 12) plan-design-review pass: 6/10 → 9/10. Added interaction-states table (6-state PrimeVue spec matching ReceiptCreateUnifiedView.vue toast conventions), design-system inheritance clause (DESIGN.md/TOKENS/ CSS_PATTERNS), a11y one-liner, week-3 subtask for enum+labels extraction from VFP source. Cloned gitea.romfast.ro/romfast/vfp_roaauto and wrote a 300+ line ground-truth audit (docs/service-auto/tabele-service-auto.md) covering: DEV_ORDL parent/child inheritance with NOM_LUCRARI, DEV_TIP_DEVIZ enum (1-7, including user-confirmed Productie/ Constatare for 6/7), dev_adauga_lucrare v1 (pack_devize, 12+1 params, body visible in packages.sql) vs v2 (pack_auto, 17+1, body only in MARIUSM_AUTO), and the pack_sesiune.dev_idLucrare trigger/session-state pattern (ROA ERP convention from pre- RETURNING era, not a bug). Plan Correction 12 added: the prototype needs a ~40-line SP doing two INSERTs parent→child with RETURNING id_lucrare (bypassing pack_sesiune), targeting dev_ordl via nom_lucrari. Original assumption of a 4-param new SP was wrong. Week 3 subtask updated to reference the audit draft. ACT/RUL untouched at creation time — Correction 3 still valid. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-06-05 09:37:09 +00:00
Claude Agent	f79a03062c	docs(service-auto): add v2 design doc for learning prototype CEO review output: rethink of scrapped 2026-04-10 product plan, reframed as learning probe + talent-risk hedge. 24-week timeline, 1 screen + 1 SP + MARIUSM_AUTO test schema. Scope wall absolute: zero touch on 4 existing VFP clients. Document includes 8 review corrections vs first draft: register_server over MarciusLabPool, SP targets dedicated orders table not ACT ledger, explicit connection.commit(), Oracle error code range classification, file handler logging, kill criterion, integration checkpoints at weeks 6+12, template+decision-log as week-24 deliverable. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-06-05 09:37:09 +00:00

Author

SHA1

Message

Date

Claude Agent

6aceb85bf1

docs(service-auto): design review + VFP audit (Correction 12)

plan-design-review pass: 6/10 → 9/10. Added interaction-states table
(6-state PrimeVue spec matching ReceiptCreateUnifiedView.vue toast
conventions), design-system inheritance clause (DESIGN.md/TOKENS/
CSS_PATTERNS), a11y one-liner, week-3 subtask for enum+labels
extraction from VFP source.

Cloned gitea.romfast.ro/romfast/vfp_roaauto and wrote a 300+ line
ground-truth audit (docs/service-auto/tabele-service-auto.md)
covering: DEV_ORDL parent/child inheritance with NOM_LUCRARI,
DEV_TIP_DEVIZ enum (1-7, including user-confirmed Productie/
Constatare for 6/7), dev_adauga_lucrare v1 (pack_devize, 12+1
params, body visible in packages.sql) vs v2 (pack_auto, 17+1,
body only in MARIUSM_AUTO), and the pack_sesiune.dev_idLucrare
trigger/session-state pattern (ROA ERP convention from pre-
RETURNING era, not a bug).

Plan Correction 12 added: the prototype needs a ~40-line SP doing
two INSERTs parent→child with RETURNING id_lucrare (bypassing
pack_sesiune), targeting dev_ordl via nom_lucrari. Original
assumption of a 4-param new SP was wrong. Week 3 subtask updated
to reference the audit draft. ACT/RUL untouched at creation time
— Correction 3 still valid.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-06-05 09:37:09 +00:00

Claude Agent

f79a03062c

docs(service-auto): add v2 design doc for learning prototype

CEO review output: rethink of scrapped 2026-04-10 product plan,
reframed as learning probe + talent-risk hedge. 24-week timeline,
1 screen + 1 SP + MARIUSM_AUTO test schema. Scope wall absolute:
zero touch on 4 existing VFP clients.

Document includes 8 review corrections vs first draft:
register_server over MarciusLabPool, SP targets dedicated orders
table not ACT ledger, explicit connection.commit(), Oracle error
code range classification, file handler logging, kill criterion,
integration checkpoints at weeks 6+12, template+decision-log as
week-24 deliverable.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-06-05 09:37:09 +00:00

2 Commits