docs(guides): import-and-export 按实测改写——导入向导在列表视图而非 Setup → Data，Salesforce 迁移标注未落地 (#763)

.changeset/import-export-guide-measured.md

@@ -0,0 +1,63 @@
+---
+'hotcrm': patch
+---
+
+Docs: restate the Import & Export guide against what the app actually does, and
+make it agree with `importing-your-data.mdx` instead of contradicting it.
+
+The page routed almost every step through a `Setup → Data → …` menu path. The
+Setup app's groups are Overview, Apps, People & Organization, Access Control,
+Approvals, Configuration, Diagnostics, Integrations and Advanced — there is no
+**Data** group and no **Privacy** group, so `Setup → Data → Import Wizard`,
+`→ Migrate from Salesforce`, `→ Migrate from HubSpot`, `→ Scheduled Exports` and
+`Setup → Privacy → Data Subject Requests` all sent the reader looking for
+screens that do not exist.
+
+Measured, and restated:
+
+- **The import wizard is real — it just lives somewhere else.** It is on the
+  object's own list view (**Import** in the toolbar): Upload → Mapping →
+  Preview, with drag-and-drop CSV / Excel or paste, a generated column template,
+  auto-matched columns with a confidence reading, write mode (*Always create
+  new* / *Update existing (skip if no match)* / *Update if matched, else
+  create*) plus **Match on**, a **Validate data** pass that writes nothing, a
+  background job with progress for large files, and **History** with **Undo
+  import**. It reads HotCRM's own saved mappings — the wizard lists the mappings
+  whose `targetObject` is the object you are on, so `crm_account_import`,
+  `crm_contact_import` and `crm_lead_import` are the same three the API path in
+  the `guides/importing-your-data` page names. The two pages
+  now describe one feature from two ends and link to each other.
+- **No completion email.** Progress is on screen and the run lands in History;
+  the platform email service is outbound-only and delivers nothing until a
+  deployment configures a transport.
+- **External IDs replaced by matching keys.** No HotCRM object carries an
+  `external_id` field, and a named mapping is a strict projection that drops
+  every column it does not declare, so the advice to add one was inert. The
+  idempotency it promised is real and comes from each mapping's business key
+  (Account Name, Email, Email). Also corrected: *Update only* skips rather than
+  fails, matching is on fields you choose rather than "Email / External ID",
+  and the 50 MB upload cap does not exist (the real ceiling is 50,000 rows per
+  job).
+- **Salesforce migration marked *(not shipped yet)***, design intent and object
+  map kept. There is no migration wizard, no connector, and no action that
+  reads Salesforce — every mention of Salesforce in `src/` is a comment citing
+  it as a design reference. There is no outbound OAuth either: Setup's **OAuth
+  Applications** registers clients calling *into* ObjectStack, the opposite
+  direction. Noted honestly that Salesforce migration is not on the roadmap,
+  while **HubSpot import is** a named roadmap item — the two are not in the
+  same state.
+- **Export restated on measurement**: the account, contact, lead and opportunity
+  list views declare `exportOptions: ['csv', 'xlsx']`, the export carries the
+  view's filter / sort / columns, and the same route is callable directly.
+  `allowExport` is the opt-in gate (unset denies; `viewAllRecords` /
+  `modifyAllRecords` do not substitute), with the per-role grants tabulated.
+  Cases hold an export grant with no list-view surface behind it, which is
+  called out rather than papered over. Scheduled warehouse export marked
+  *(not shipped yet)*.
+- **GDPR section marked *(not shipped yet)*** and replaced with what can be done
+  today — delete the contact (attendee rows go, meetings stay), delete the
+  account (contacts cascade), clear identifying fields by hand. No Privacy
+  group, no data-subject-request object, no anonymise action, no audit
+  certificate (the audit capability is not among the ones this app enables).
+
+zh-Hans and zh-Hant pages updated with the same content.