Commit `d5dec7be8a`

d5dec7be8a88165dcb81c88c6fa23aa7e7360a03

parent: 073fca4510

Unsigned

cmc <hello@cleberg.net> · 2026-07-16 15:04 UTC

docs: bucket the Phase 3 backlog into semantic versions

Layout: unified · split

ROADMAP.md +67 −8

		@@ -134,6 +134,28 @@ GraphQL mutation. Treat that boundary as explicit rather than half-building it.
134	134	Unlike Phases 1 and 2, this is not one shippable thing. It is several, and they
135	135	are sized very differently — measure before committing to one.
136	136
	137	### Release plan
	138
	139	Hutch is an app with a `MARKETING_VERSION`, not a library with an API contract,
	140	so "breaking change" does not apply. These buckets track user-visible scale.
	141
	142	\| Version \| Contents \| Why here \|
	143	\| --- \| --- \| --- \|
	144	\| v3.8.1 \| SonarCloud triage; housekeeping \| No behaviour change at all \|
	145	\| v3.9.0 \| "What's cooking" ingest; doc truth-up; deploy keys \| Ships one feature, corrects the map \|
	146	\| v3.10.0 \| hub.sr.ht writes: projects, discovery, `mailingListSubscribe` \| Provisional — gated on what v3.9.0 finds \|
	147	\| v3.11.0 \| Accessibility \| Independent, device-verified \|
	148	\| v4.0.0 \| Localization with translations \| The only true re-presentation \|
	149	\| — \| Swift 6 language mode; cache reads \| Internal; ride along, no tag \|
	150
	151	Ordering is by dependency, not size. v3.9.0 leads because it is the only item
	152	that corrects the others' inputs: the ingest's real output is a `SCOPE.md` that
	153	is true, and v3.10.0 rests entirely on one unverified sentence in a blog post.
	154	Do not commit v3.10.0's number until the SDL has been read — the bucket may turn
	155	out to be empty, which is the point of sequencing it second.
	156
	157	`KeychainHelper` is deliberately unbucketed; see the SonarCloud hotspots below.
	158
137	159	### API features — done (v3.8.0)
138	160
139	161	- ~~`uploadArtifact` / `deleteArtifact`~~ — artifacts were read-only.
		@@ -152,7 +174,7 @@ Three of the six planned. The other three did not survive contact:
152	174	on judgement — see [SCOPE.md](SCOPE.md) for the reasoning, so they do not get
153	175	re-proposed.
154	176
155		### Localization
	177	### Localization — v4.0.0, and only with translations
156	178
157	179	The project sets `LOCALIZATION_PREFERS_STRING_CATALOGS = YES` but ships no
158	180	string catalog, so every user-facing string is hardcoded English. Roughly 634
		@@ -164,12 +186,19 @@ for users until translations exist. It is groundwork, and it is the largest diff
164	186	in the roadmap — it touches nearly every view, with the regression risk that
165	187	implies.
166	188
167		### Accessibility
	189	That combination is why this is bucketed at v4.0.0 *bundled with at least one
	190	real translation*, rather than shipped alone. An English-only catalog would earn
	191	the major number on regression risk while delivering nothing — the wrong trade.
	192	Hold the catalog until a translation lands. If it ever ships unbundled, it is
	193	groundwork and belongs in a quiet minor, not a 4.0.
	194
	195	### Accessibility — v3.11.0
168	196
169	197	Labels and hints appear in 17 of 89 view files. Mechanical and low-risk, but it
170	198	cannot be verified from a build — it needs VoiceOver driven on a device.
	199	Independent of every other bucket, so it can move if a device pass is convenient.
171	200
172		### SonarCloud backlog
	201	### SonarCloud backlog — v3.8.1
173	202
174	203	51 open issues: 0 bugs, 0 vulnerabilities, 51 code smells, plus 3 security
175	204	hotspots. The headline number is misleading, so trust the breakdown before
		@@ -193,7 +222,11 @@ The 3 hotspots are the part actually worth thought:
193	222	`kSecAttrAccessibleWhenUnlockedThisDeviceOnly` with no
194	223	`SecAccessControl`, so it does not require biometric or passcode
195	224	authentication to read. That is a genuine product decision — should a stolen,
196		unlocked phone hand over a sr.ht token? — not a lint nit.
	225	unlocked phone hand over a sr.ht token? — not a lint nit. **Unbucketed on
	226	purpose:** adding `SecAccessControl` changes what a user must do to read their
	227	own token, so it needs a decision first. If the answer is yes, it is a minor
	228	bump of its own — a visible auth change should not hide inside a feature
	229	release.
197	230	- `ReadmeView:1922` (LOW) — unrestricted WebView navigation. Probably a false
198	231	positive: `isAllowedReadmeNavigationURL` enforces a scheme allowlist. Verify,
199	232	then annotate.
		@@ -201,7 +234,12 @@ The 3 hotspots are the part actually worth thought:
201	234	Query it with:
202	235	`https://sonarcloud.io/api/issues/search?componentKeys=zerolabsco_hutch&resolved=false`
203	236
204		### Ingest "What's cooking on SourceHut?"
	237	This is a patch because nothing executes differently afterwards. The 35 hardcoded-URI
	238	issues are resolved as Won't Fix in SonarCloud's web UI — not a commit at all — and
	239	the rest is three comments and one annotation. If it produces a diff that changes a
	240	runtime path, something has gone wrong.
	241
	242	### Ingest "What's cooking on SourceHut?" — v3.9.0
205	243
206	244	sr.ht posts a quarterly update to `~sircmpwn/sr.ht-announce`, mirrored at
207	245	<https://sourcehut.org/blog/>. Nothing in Hutch tracks it, so the API grows and
		@@ -227,7 +265,25 @@ flags two openings:
227	265	Start from Q1 2026 forward — that is roughly when the current `Docs/API` dumps
228	266	were captured.
229	267
230		### Swift 6 language mode
	268	Research does not ship, so v3.9.0 pairs the ingest with deploy keys — the one
	269	self-contained feature it has already surfaced and that the SDL confirms exists.
	270	That gives the release something a user can see. Everything else the ingest turns
	271	up gets filed, not built, and hub.sr.ht gets its own bucket below.
	272
	273	### hub.sr.ht writes — v3.10.0, provisional
	274
	275	Everything here rests on a single sentence in the Q2 2026 post: that hub.sr.ht
	276	gained a writable GraphQL API. If true, three things unblock at once —
	277	project writes (Hutch's projects are read-only), discovery (which `SCOPE.md`
	278	rules out on the grounds hub has no public API), and `mailingListSubscribe`,
	279	which Phase 1 declined for exactly that reason.
	280
	281	All three live or die on the same unverified claim, which is why this is
	282	sequenced after the ingest rather than planned now. Read
	283	`api/graph/schema.graphqls` in `hub.sr.ht` before committing the version number.
	284	The bucket may be empty.
	285
	286	### Swift 6 language mode — no release of its own
231	287
232	288	The project builds in Swift 5 language mode with
233	289	`SWIFT_DEFAULT_ACTOR_ISOLATION = MainActor`. Moving to Swift 6 is blocked on
		@@ -243,14 +299,17 @@ concurrency diagnostics that are warnings today and errors there:
243	299	methods, as in `HomeViewModel.loadDashboard` and
244	300	`NotificationPreferencesViewModel.load`.
245	301
246		### Cache reads that bypass the client
	302	### Cache reads that bypass the client — no release of its own
247	303
248	304	`BuildListViewModel`, `RepositoryListViewModel`, and `PasteService` still read
249	305	`client.responseCache` directly, each falling back across two different cache
250	306	keys. That predates `APICacheKeys` and should be folded into `cachedPayload`,
251	307	which already consults the persistent cache before the memory layer.
252	308
253		## Housekeeping
	309	Like Swift 6 above, this is internal and rides along with whatever release
	310	already touches that area. Neither justifies a tag.
	311
	312	## Housekeeping — v3.8.1
254	313
255	314	- `Hutch/Hutch/App/AccountSession.swift` sits in a stray nested directory;
256	315	`Hutch/HutchTests/` is empty.

Commit d5dec7be8a

Commit `d5dec7be8a`