@dan-ai-studio/dshopencodego 0.1.19 → 0.1.21

package/README.en.md

@@ -162,7 +162,7 @@ The `engines.dsh` field is informational for readers and package managers; **DSH
 Things to keep in mind when maintaining this:
 
 - **Required versus optional peers**: six of the eighteen are marked `optional` (`dsh-api-remotes`, `dsh-client-locale`, `dsh-client-store`, `dsh-client-ui-model-selection`, `dsh-client-ui-settings`, `dsh-client-ui-slots`). The twelve that actually block loading are `cordis`, `dsh-attachment`, `dsh-brand`, `dsh-client-ui-conversation`, `dsh-client-ui-renderer`, `dsh-credentials`, `dsh-fs`, `dsh-launch-environment`, `dsh-llm`, `dsh-timeout`, `dsh-typert-protocol`, and `dsh-typert-registry`. `dsh-llm` is the heaviest coupling — more than twenty imports across the source.
-- **Declared range ≠ verified range**: `0.1.7-alpha.1` and `alpha.2` fall inside the declaration but were never verified here. What was verified is `0.1.7-rc.1`, `0.1.7-rc.2` (daily use), `0.2.0-rc.1` (the dev-dependency baseline: gate verdict + typecheck + full suite), and `0.2.0-rc.2` (dev-dependency baseline + `prepareCall` generation freeze + pi-ai delta-parse patch port).
+- **Declared range ≠ verified range**: `0.1.7-alpha.1` and `alpha.2` fall inside the declaration but were never verified here. What was verified is `0.1.7-rc.1`, `0.1.7-rc.2` (daily use), `0.2.0-rc.1` (the dev-dependency baseline: gate verdict + typecheck + full suite), and `0.2.0-rc.2` (dev-dependency baseline + `prepareCall` generation freeze).
 - **The upper bound is a tracking line**: `<0.3.0` means that the moment DSH ships `0.3.0`, this plugin must ship a new release in the same window or every user loses the plugin. The current range spans both `0.1.7` and `0.2.0` (`0.1.8` and `0.1.9` were never published, hence two segments). After changing a range, `npm test` is the verification (local mock gateway, no network, no tokens) — ranges follow what the interfaces declare, and are never "measured" with live probes.
 - **Peer coverage is now guarded both ways**: `dsh-typert-registry`, `dsh-client-ui-conversation`, and `dsh-client-ui-renderer` were imported by the source but absent from `peerDependencies`, invisible to the gate; they are declared now, and `tests/peer-coverage.spec.ts` asserts both directions — anything the source imports must be declared, and anything declared but unreached must be removed (`dsh-settings` went that way: the settings form is actually provided by `dsh-client-ui-settings`).
 - **Exemptions are per profile**: only a profile that ran `dsh plugin allow-version` has one, and no record means the plugin must stay inside the declared range. `dsh plugin version-exemptions` lists a profile's exemptions.
@@ -179,6 +179,7 @@ Things to keep in mind when maintaining this:
 | "The settings write was rejected" | Usually a concurrent writer (another window or process) holding the same profile. The plugin retries once automatically; if it still fails the view has reloaded — click again. |
 | The usage button shows "unavailable" | An endpoint or credential change invalidated the previous reading. Fix the configuration and press Retry. It never shows `0` in place of a missing reading. |
 | npm publish fails with `ENEEDAUTH` | OIDC did not match: check the Trusted Publisher's **workflow filename** (must be exactly `publish-npm.yml`, case-sensitive) and the organization/repository fields; run on GitHub-hosted runners with `id-token: write`. A `repository.url` that does not match the GitHub repository is rejected as well. |
+| The whole DSH process stalls during ultra-long tool calls (multi-MB arguments) | Inherent to pi-ai 0.87.1's per-delta re-parsing (issue #9265 / PR #9461). This plugin does not port DSH's consumer-side patch (ADR-0005); the fix arrives with an upstream release and a pi-ai upgrade. |
 
 ## Development
 

package/README.md

@@ -162,7 +162,7 @@ API Key 通过 Harness 凭证库提供（引用名 `OPENCODE_GO_API_KEY`），
 维护时需要注意：
 
 - **必选与可选之分**：18 个 peer（17 个 `dsh-*` 加框架 `cordis`）中 6 个标了 `optional`（`dsh-api-remotes`、`dsh-client-locale`、`dsh-client-store`、`dsh-client-ui-model-selection`、`dsh-client-ui-settings`、`dsh-client-ui-slots`）；真正卡住加载的是 12 个——`cordis`、`dsh-attachment`、`dsh-brand`、`dsh-client-ui-conversation`、`dsh-client-ui-renderer`、`dsh-credentials`、`dsh-fs`、`dsh-launch-environment`、`dsh-llm`、`dsh-timeout`、`dsh-typert-protocol`、`dsh-typert-registry`。其中 `dsh-llm` 是最重的一处耦合（源码 import 二十余处）。
-- **名义范围 ≠ 实测范围**：`0.1.7-alpha.1`、`alpha.2` 落在声明范围内，但本项目没有验证过；实际验证过的是 `0.1.7-rc.1`、`0.1.7-rc.2`（日常使用）、`0.2.0-rc.1`（开发依赖基线：门禁判定 + typecheck + 全量用例）与 `0.2.0-rc.2`（开发依赖基线 + `prepareCall` 代次冻结 + pi-ai 增量解析补丁移植）。
+- **名义范围 ≠ 实测范围**：`0.1.7-alpha.1`、`alpha.2` 落在声明范围内，但本项目没有验证过；实际验证过的是 `0.1.7-rc.1`、`0.1.7-rc.2`（日常使用）、`0.2.0-rc.1`（开发依赖基线：门禁判定 + typecheck + 全量用例）与 `0.2.0-rc.2`（开发依赖基线 + `prepareCall` 代次冻结）。
 - **上界就是跟版线**：`<0.3.0` 意味着 DSH 一旦发布 `0.3.0`，本插件必须同批发新版，否则所有用户加载失败。当前范围跨 `0.1.7` 与 `0.2.0` 两代（`0.1.8`、`0.1.9` 从未发布，所以写成两段而不是一条）。调整范围后用 `npm test` 验证即可（本地 mock 网关，零网络零 token）——范围以接口声明为准，不要用线上探测来"测"出一个范围。
 - **peer 覆盖已双向守住**：`dsh-typert-registry`、`dsh-client-ui-conversation`、`dsh-client-ui-renderer` 曾被源码 import 却未声明、门禁管不到，现已补进 `peerDependencies`；`tests/peer-coverage.spec.ts` 双向断言——源码用到的必须声明，声明了却没人用的必须删（`dsh-settings` 就这样被移除：设置表单实际由 `dsh-client-ui-settings` 提供）。
 - **豁免是按 profile 记的**：只有显式执行过 `dsh plugin allow-version` 的 profile 才享有豁免；没有记录就代表必须落在声明范围内。用 `dsh plugin version-exemptions` 可查看当前 profile 的豁免。
@@ -179,6 +179,7 @@ API Key 通过 Harness 凭证库提供（引用名 `OPENCODE_GO_API_KEY`），
 | 设置页出现「设置写入被拒绝」 | 通常是与另一个窗口/进程并发写同一 profile；插件会自动重试一次，仍失败时界面已重载最新状态，再点一次即可。 |
 | 用量按钮显示「不可用」 | 端点或凭证变化导致旧读数作废；配置正确后点「重试」。数值不会以 0 冒充。 |
 | npm 发布报 `ENEEDAUTH` | OIDC 没匹配上：核对包设置里 Trusted Publisher 的 **workflow 文件名**（必须恰为 `publish-npm.yml`、大小写敏感）与组织/仓库字段；确认跑在 GitHub 托管 runner 上且工作流带 `id-token: write`。`package.json` 的 `repository.url` 与仓库不符也会被拒。 |
+| 超长工具调用（多 MB 参数）期间整个 DSH 进程卡顿 | 上游 pi-ai 0.87.1 逐 delta 重解析的固有开销（issue #9265 / PR #9461）。本插件不移植 DSH 的消费者侧补丁（ADR-0005）；修复随上游发布后升级 pi-ai 解决。 |
 
 ## 开发
 

package/icon.svg

@@ -0,0 +1,11 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64" role="img" aria-label="OpenCode Go">
+  <defs>
+    <linearGradient id="bg" x1="0" y1="0" x2="1" y2="1">
+      <stop offset="0" stop-color="#1f6feb"/>
+      <stop offset="1" stop-color="#0d4196"/>
+    </linearGradient>
+  </defs>
+  <rect x="1.5" y="1.5" width="61" height="61" rx="14" fill="url(#bg)" stroke="#0a2f6b" stroke-width="3"/>
+  <path d="M20 21.5 33.5 32 20 42.5" fill="none" stroke="#ffffff" stroke-width="5.5" stroke-linecap="round" stroke-linejoin="round"/>
+  <path d="M33 21.5 46.5 32 33 42.5" fill="none" stroke="#9ecbff" stroke-width="5.5" stroke-linecap="round" stroke-linejoin="round"/>
+</svg>

package/lib/client.js

@@ -462,7 +462,7 @@ if (!document.querySelector("style[data-plugin-css=" + JSON.stringify(id) + "]")
   style.textContent = ".d7lz0W_root{color:var(--dsw-alias-label-primary);flex-direction:column;gap:1.25rem;display:flex}.d7lz0W_title{margin:0;font-size:1.05rem}.d7lz0W_hint{color:var(--dsw-alias-label-secondary);margin:0;font-size:.8rem}.d7lz0W_warn{color:var(--dsw-alias-state-warn-primary);margin:0;font-size:.8rem}.d7lz0W_ok{color:var(--dsw-alias-state-success-primary);font-size:.8rem}.d7lz0W_block{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-layer-1);border-radius:.6rem;flex-direction:column;gap:.5rem;padding:.85rem;display:flex}.d7lz0W_row{flex-wrap:wrap;align-items:center;gap:.5rem;display:flex}.d7lz0W_row>strong{margin-right:auto}.d7lz0W_input{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-base);min-width:10rem;color:var(--dsw-alias-label-primary);font:inherit;border-radius:.35rem;flex:14rem;padding:.35rem .5rem;font-size:.85rem}.d7lz0W_button{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-layer-2);color:var(--dsw-alias-label-primary);font:inherit;cursor:pointer;border-radius:.35rem;padding:.35rem .7rem;font-size:.8rem}.d7lz0W_button:disabled{color:var(--dsw-alias-state-idle-primary);cursor:default}.d7lz0W_number{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-base);width:5rem;color:var(--dsw-alias-label-primary);font:inherit;border-radius:.35rem;padding:.25rem .4rem;font-size:.8rem}.d7lz0W_select{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-base);width:auto;max-width:100%;color:var(--dsw-alias-label-primary);font:inherit;border-radius:.35rem;padding:.25rem .4rem;font-size:.8rem}.d7lz0W_list{flex-direction:column;display:flex}.d7lz0W_item{border-top:1px solid var(--dsw-alias-border-l1);align-items:flex-start;gap:.5rem;padding:.45rem .15rem;display:flex}.d7lz0W_item input[type=checkbox]{accent-color:var(--dsw-alias-brand-primary);margin-top:.15rem}.d7lz0W_itemBody{flex-direction:column;gap:.15rem;min-width:0;display:flex}.d7lz0W_itemMain{flex-wrap:wrap;align-items:baseline;gap:.6rem;display:flex}.d7lz0W_itemMeta{color:var(--dsw-alias-label-secondary);flex-wrap:wrap;gap:.75rem;font-size:.75rem;display:flex}.d7lz0W_itemModalities{color:var(--dsw-alias-label-secondary);font-size:.75rem}.d7lz0W_name{color:var(--dsw-alias-label-primary)}.d7lz0W_mono{color:var(--dsw-alias-label-secondary);font-family:ui-monospace,SFMono-Regular,Menlo,monospace}.d7lz0W_badge{color:var(--dsw-alias-brand-primary)}";
   document.head.appendChild(style);
 }
-var section_default = { "row": "d7lz0W_row", "itemModalities": "d7lz0W_itemModalities", "input": "d7lz0W_input", "button": "d7lz0W_button", "list": "d7lz0W_list", "root": "d7lz0W_root", "item": "d7lz0W_item", "title": "d7lz0W_title", "itemMeta": "d7lz0W_itemMeta", "itemBody": "d7lz0W_itemBody", "itemMain": "d7lz0W_itemMain", "badge": "d7lz0W_badge", "ok": "d7lz0W_ok", "number": "d7lz0W_number", "name": "d7lz0W_name", "mono": "d7lz0W_mono", "select": "d7lz0W_select", "hint": "d7lz0W_hint", "block": "d7lz0W_block", "warn": "d7lz0W_warn" };
+var section_default = { "itemBody": "d7lz0W_itemBody", "badge": "d7lz0W_badge", "item": "d7lz0W_item", "itemModalities": "d7lz0W_itemModalities", "row": "d7lz0W_row", "hint": "d7lz0W_hint", "root": "d7lz0W_root", "itemMeta": "d7lz0W_itemMeta", "block": "d7lz0W_block", "mono": "d7lz0W_mono", "button": "d7lz0W_button", "list": "d7lz0W_list", "warn": "d7lz0W_warn", "ok": "d7lz0W_ok", "select": "d7lz0W_select", "itemMain": "d7lz0W_itemMain", "number": "d7lz0W_number", "name": "d7lz0W_name", "title": "d7lz0W_title", "input": "d7lz0W_input" };
 
 // src/client/Section.tsx
 var import_jsx_runtime = require("react/jsx-runtime");
@@ -963,7 +963,7 @@ if (!document.querySelector("style[data-plugin-css=" + JSON.stringify(id2) + "]"
   style.textContent = ".mMNkkW_root{align-items:center;display:inline-flex;position:relative}.mMNkkW_trigger{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-layer-1);color:var(--dsw-alias-label-secondary);font:inherit;cursor:pointer;white-space:nowrap;border-radius:999px;padding:.35rem .6rem;font-size:.8rem;line-height:1}.mMNkkW_trigger:hover{color:var(--dsw-alias-label-primary);border-color:var(--dsw-alias-border-l2)}.mMNkkW_panel{z-index:20;border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-overlay);width:18rem;color:var(--dsw-alias-label-primary);border-radius:.6rem;flex-direction:column;gap:.6rem;padding:.75rem .85rem;display:flex;position:absolute;bottom:calc(100% + .5rem);right:0;box-shadow:0 8px 24px #0000002e}.mMNkkW_window{flex-direction:column;gap:.25rem;display:flex}.mMNkkW_row{justify-content:space-between;gap:.5rem;font-size:.85rem;display:flex}.mMNkkW_hint{color:var(--dsw-alias-label-secondary);margin:0;font-size:.75rem}.mMNkkW_progress{width:100%;height:.35rem;accent-color:var(--dsw-alias-brand-primary)}.mMNkkW_limited{color:var(--dsw-alias-state-warn-primary);font-size:.75rem}.mMNkkW_warning{border:1px solid var(--dsw-alias-state-warn-primary);color:var(--dsw-alias-label-primary);border-radius:.4rem;padding:.4rem .5rem;font-size:.75rem}.mMNkkW_retry{border:1px solid var(--dsw-alias-border-l1);background:var(--dsw-alias-bg-layer-1);color:var(--dsw-alias-label-primary);font:inherit;cursor:pointer;border-radius:.35rem;align-self:flex-start;padding:.25rem .5rem;font-size:.75rem}.mMNkkW_local{border-top:1px solid var(--dsw-alias-border-l1);flex-direction:column;gap:.2rem;padding-top:.5rem;display:flex}";
   document.head.appendChild(style);
 }
-var pill_default = { "local": "mMNkkW_local", "row": "mMNkkW_row", "retry": "mMNkkW_retry", "panel": "mMNkkW_panel", "warning": "mMNkkW_warning", "window": "mMNkkW_window", "hint": "mMNkkW_hint", "root": "mMNkkW_root", "progress": "mMNkkW_progress", "limited": "mMNkkW_limited", "trigger": "mMNkkW_trigger" };
+var pill_default = { "limited": "mMNkkW_limited", "retry": "mMNkkW_retry", "window": "mMNkkW_window", "row": "mMNkkW_row", "hint": "mMNkkW_hint", "panel": "mMNkkW_panel", "warning": "mMNkkW_warning", "local": "mMNkkW_local", "root": "mMNkkW_root", "progress": "mMNkkW_progress", "trigger": "mMNkkW_trigger" };
 
 // src/client/UsagePill.tsx
 var import_jsx_runtime2 = require("react/jsx-runtime");

package/lib/index.js

@@ -1417,6 +1417,14 @@ function withModelLimit(model, limits) {
     maxTokens: limit.maxTokens ?? model.maxTokens
   };
 }
+function stopSequencesFor(model, stop) {
+  if (stop === void 0) return void 0;
+  if (model.api === "openai-completions") return stop;
+  throw new LlmError8(
+    `opencode-go model "${model.id}" speaks ${model.api}, which has no stop-sequence parameter this adapter can reach; drop GenerateOptions.stop for this model`,
+    "UNSUPPORTED_OPTION"
+  );
+}
 var OpencodeGoAdapter = class extends LlmAdapter {
   cache;
   options;
@@ -1538,12 +1546,10 @@ var OpencodeGoAdapter = class extends LlmAdapter {
   async *streamWithSnapshot(options, config, snapshot) {
     var _stack = [];
     try {
-      if (options.stop !== void 0) {
-        throw new LlmError8("dshopencodego does not support GenerateOptions.stop", "UNSUPPORTED_OPTION");
-      }
       const facts = snapshot.facts.get(options.model);
       if (facts === void 0) throw new LlmError8(`opencode-go has no model "${options.model}"`, "UNKNOWN_MODEL");
       const model = withModelLimit(toPiModel(facts, config.baseURL), config.modelLimits);
+      const stop = stopSequencesFor(model, options.stop);
       const ceiling = config.modelLimits[model.id]?.maxTokens;
       const maxTokens = options.maxTokens === void 0 || ceiling === null || ceiling === void 0 ? options.maxTokens : Math.min(options.maxTokens, ceiling);
       const apiKey = await this.options.resolveApiKey();
@@ -1579,6 +1585,10 @@ var OpencodeGoAdapter = class extends LlmAdapter {
           ...reasoning === void 0 || reasoning === "off" ? {} : { reasoning },
           ...options.temperature === void 0 ? {} : { temperature: options.temperature },
           ...maxTokens === void 0 ? {} : { maxTokens },
+          // `samplingParams` is merged into the request body as-is by the
+          // OpenAI-compatible adapters, and `stop` is Chat Completions' own
+          // stop-sequence field — the one place this control reaches the wire.
+          ...stop === void 0 ? {} : { samplingParams: { stop: [...stop] } },
           ...options.sessionId === void 0 ? {} : { sessionId: String(options.sessionId) },
           signal: watchdog.signal,
           // The gateway refuses a request without `x-opencode-session`; the
@@ -1858,7 +1868,7 @@ import { RemoteError, TypertRemoteService } from "@deepseek-ai/dsh-typert-protoc
 import { LlmError as LlmError9 } from "@deepseek-ai/dsh-llm";
 
 // src/catalog/reading.ts
-function catalogReading(snapshot, visibility, listingFailure) {
+function catalogReading(snapshot, visibility, listingFailure, now) {
   const documentQuotas = snapshot.documentQuotas;
   const rows = [
     ...[...snapshot.facts.values()].map((fact) => {
@@ -1892,7 +1902,7 @@ function catalogReading(snapshot, visibility, listingFailure) {
   const models = rows.map((row) => ({ ...row, recommended: recommended.has(row.id) }));
   const enabled = models.filter((model) => isModelEnabled(model, visibility)).length;
   return {
-    models: sortModels(models),
+    models: sortModels(models, now),
     stale: !snapshot.live,
     ...!snapshot.live && listingFailure !== void 0 ? { error: listingFailure } : {},
     fetchedAtMs: snapshot.fetchedAtMs,

package/lib/types/adapter.js

@@ -90,6 +90,31 @@ function withModelLimit(model, limits) {
         maxTokens: limit.maxTokens ?? model.maxTokens,
     };
 }
+/**
+ * Resolve one request's stop sequences against the model's protocol.
+ *
+ * The seam's stop control reaches a route only through the provider's own
+ * field, and pi-ai models none: its Anthropic options carry no stop sequences
+ * and that adapter ignores `samplingParams`, and its Responses options define
+ * no stop-sequence parameter at all. Only the OpenAI-compatible adapters merge
+ * arbitrary body fields, and `stop` is exactly the field Chat Completions
+ * defines, so a completions model can honor it. Every other protocol refuses:
+ * silently dropping the sequences would let a caller believe generation halts
+ * at a string the model never sees.
+ * @param model - the exact model this request will use.
+ * @param stop - the caller's stop sequences, when it named any.
+ * @returns the sequences to send, or undefined when the caller named none.
+ * @throws {LlmError} `UNSUPPORTED_OPTION` for a protocol with no reachable
+ *   stop-sequence parameter.
+ */
+function stopSequencesFor(model, stop) {
+    if (stop === undefined)
+        return undefined;
+    if (model.api === 'openai-completions')
+        return stop;
+    throw new LlmError(`opencode-go model "${model.id}" speaks ${model.api}, which has no stop-sequence parameter this adapter can reach;`
+        + ' drop GenerateOptions.stop for this model', 'UNSUPPORTED_OPTION');
+}
 /**
  * The single `opencode-go` route's adapter.
  *
@@ -238,13 +263,13 @@ export class OpencodeGoAdapter extends LlmAdapter {
     async *streamWithSnapshot(options, config, snapshot) {
         const env_1 = { stack: [], error: void 0, hasError: false };
         try {
-            if (options.stop !== undefined) {
-                throw new LlmError('dshopencodego does not support GenerateOptions.stop', 'UNSUPPORTED_OPTION');
-            }
             const facts = snapshot.facts.get(options.model);
             if (facts === undefined)
                 throw new LlmError(`opencode-go has no model "${options.model}"`, 'UNKNOWN_MODEL');
             const model = withModelLimit(toPiModel(facts, config.baseURL), config.modelLimits);
+            // Resolved before the first await so an unsupported protocol fails the
+            // request before any credential is read or any byte is sent.
+            const stop = stopSequencesFor(model, options.stop);
             // `defaultMaxTokens` already carries the route's ceiling for a caller that
             // names none, so only a caller's own cap travels, clamped by a configured
             // per-model ceiling when one exists.
@@ -287,6 +312,10 @@ export class OpencodeGoAdapter extends LlmAdapter {
                     ...reasoning === undefined || reasoning === 'off' ? {} : { reasoning },
                     ...options.temperature === undefined ? {} : { temperature: options.temperature },
                     ...maxTokens === undefined ? {} : { maxTokens },
+                    // `samplingParams` is merged into the request body as-is by the
+                    // OpenAI-compatible adapters, and `stop` is Chat Completions' own
+                    // stop-sequence field — the one place this control reaches the wire.
+                    ...stop === undefined ? {} : { samplingParams: { stop: [...stop] } },
                     ...options.sessionId === undefined ? {} : { sessionId: String(options.sessionId) },
                     signal: watchdog.signal,
                     // The gateway refuses a request without `x-opencode-session`; the

package/lib/types/catalog/reading.d.ts

@@ -40,7 +40,9 @@ export interface CatalogReading {
  * @param snapshot - the catalog snapshot, live or retained.
  * @param visibility - per-model switches; absent entries keep the default.
  * @param listingFailure - the failure message when the listing failed.
+ * @param now - the clock the "newly shipped" ordering reads; injected so a
+ *   reading is a pure function of its inputs rather than of the wall clock.
  * @returns the reading, with unconfigurable ids kept visible so the page can
  *   explain them instead of hiding them.
  */
-export declare function catalogReading(snapshot: CatalogSnapshot, visibility: Readonly<Record<string, boolean>>, listingFailure?: string): CatalogReading;
+export declare function catalogReading(snapshot: CatalogSnapshot, visibility: Readonly<Record<string, boolean>>, listingFailure?: string, now?: number): CatalogReading;

package/lib/types/catalog/reading.js

@@ -16,10 +16,12 @@ import { goQuotaFor } from "../go-limits.js";
  * @param snapshot - the catalog snapshot, live or retained.
  * @param visibility - per-model switches; absent entries keep the default.
  * @param listingFailure - the failure message when the listing failed.
+ * @param now - the clock the "newly shipped" ordering reads; injected so a
+ *   reading is a pure function of its inputs rather than of the wall clock.
  * @returns the reading, with unconfigurable ids kept visible so the page can
  *   explain them instead of hiding them.
  */
-export function catalogReading(snapshot, visibility, listingFailure) {
+export function catalogReading(snapshot, visibility, listingFailure, now) {
     const documentQuotas = snapshot.documentQuotas;
     const rows = [
         ...[...snapshot.facts.values()].map(fact => {
@@ -55,7 +57,7 @@ export function catalogReading(snapshot, visibility, listingFailure) {
     const models = rows.map(row => ({ ...row, recommended: recommended.has(row.id) }));
     const enabled = models.filter(model => isModelEnabled(model, visibility)).length;
     return {
-        models: sortModels(models),
+        models: sortModels(models, now),
         stale: !snapshot.live,
         ...!snapshot.live && listingFailure !== undefined ? { error: listingFailure } : {},
         fetchedAtMs: snapshot.fetchedAtMs,

package/locale/en.json

@@ -0,0 +1,6 @@
+{
+  "meta": {
+    "title": "OpenCode Go",
+    "description": "OpenCode Go provider for DeepSeek Harness: live gateway catalog, per-conversation session header, usage"
+  }
+}

package/locale/zh.json

@@ -0,0 +1,6 @@
+{
+  "meta": {
+    "title": "OpenCode Go",
+    "description": "DeepSeek Harness 的 OpenCode Go 提供方：实时网关目录、按对话的会话头、用量读数"
+  }
+}

package/package.json

@@ -1,9 +1,10 @@
 {
   "name": "@dan-ai-studio/dshopencodego",
-  "version": "0.1.19",
+  "version": "0.1.21",
   "description": "OpenCode Go provider for DeepSeek Harness: live gateway catalog, per-conversation session header, usage",
   "type": "module",
   "license": "MIT",
+  "icon": "./icon.svg",
   "repository": {
     "type": "git",
     "url": "git+https://github.com/dan-ai-studio/dshopencodego.git"
@@ -27,12 +28,14 @@
       "types": "./lib/types/client/index.d.ts",
       "default": "./lib/client.js"
     },
+    "./locale/*.json": "./locale/*.json",
     "./package.json": "./package.json"
   },
   "files": [
     "lib",
-    "scripts/patch-pi-ai.mjs",
     "cordis.patch.yml",
+    "icon.svg",
+    "locale",
     "README.md",
     "README.en.md",
     "LICENSE"
@@ -60,16 +63,15 @@
     "test": "vitest run",
     "test:coverage": "vitest run --coverage",
     "test:watch": "vitest",
-    "postinstall": "node scripts/patch-pi-ai.mjs",
     "prepare": "npm run build",
     "prepack": "npm run build"
   },
   "dependencies": {
-    "@deepseek-ai/schemastery": "^3.18.3",
-    "@earendil-works/pi-ai": "0.87.1"
+    "@deepseek-ai/schemastery": "~3.18.5-alpha.1",
+    "@earendil-works/pi-ai": "~1.0.2"
   },
   "peerDependencies": {
-    "@deepseek-ai/cordis": "4.0.2 || 4.0.3 || 4.0.4",
+    "@deepseek-ai/cordis": "4.0.2 || 4.0.3 || 4.0.4 || 4.0.5-alpha.1",
     "@deepseek-ai/dsh-api-remotes": ">=0.1.7-alpha.1 <0.2.0 || >=0.2.0-rc.1 <0.3.0",
     "@deepseek-ai/dsh-attachment": ">=0.1.7-alpha.1 <0.2.0 || >=0.2.0-rc.1 <0.3.0",
     "@deepseek-ai/dsh-brand": ">=0.1.7-alpha.1 <0.2.0 || >=0.2.0-rc.1 <0.3.0",
@@ -109,29 +111,28 @@
     }
   },
   "devDependencies": {
-    "@deepseek-ai/cordis": "4.0.4",
-    "@deepseek-ai/dsh-api-remotes": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-attachment": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-brand": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-locale": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-store": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-ui-conversation": "^0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-ui-model-selection": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-ui-renderer": "^0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-ui-settings": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-client-ui-slots": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-credentials": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-fs": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-invariants": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-launch-environment": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-llm": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-sandbox": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-scope": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-session": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-settings": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-timeout": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-typert-protocol": "0.2.0-rc.2",
-    "@deepseek-ai/dsh-typert-registry": "^0.2.0-rc.2",
+    "@deepseek-ai/cordis": "4.0.5-alpha.1",
+    "@deepseek-ai/dsh-api-remotes": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-attachment": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-brand": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-locale": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-store": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-ui-conversation": "^0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-ui-model-selection": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-ui-renderer": "^0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-ui-settings": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-client-ui-slots": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-credentials": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-fs": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-launch-environment": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-llm": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-sandbox": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-scope": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-session": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-settings": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-timeout": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-typert-protocol": "0.2.1-alpha.2",
+    "@deepseek-ai/dsh-typert-registry": "^0.2.1-alpha.2",
     "@testing-library/dom": "^10.4.1",
     "@testing-library/react": "^16.3.0",
     "@types/node": "^22.20.0",

package/scripts/patch-pi-ai.mjs

@@ -1,115 +0,0 @@
-/**
- * Port of DSH's `patches/@earendil-works__pi-ai@0.87.1.patch` for npm installs.
- *
- * Upstream re-parses the whole accumulated tool-call argument JSON on every
- * stream delta (`O(n^2)` on multi-megabyte arguments and stalls the event
- * loop). DSH removes those per-delta parses via pnpm `patchedDependencies`;
- * this package is npm-installed, so the same removals are applied here,
- * idempotently, to the installed copy.
- *
- * Only the per-delta occurrences are removed — each is anchored on its
- * distinctive preceding accumulation line, exactly matching the upstream
- * hunks. The final parses (toolcall_end / done handlers, which strip the
- * scratch buffers so replay carries parsed arguments) are kept.
- *
- * The adapter only reads delta strings plus the finalized arguments (see
- * `src/conversion/stream.ts`), so dropping the intermediate `arguments`
- * assignments changes no wire behaviour — until `toolcall_end` the partials
- * stay `{}`, exactly like the patched DSH runtime.
- *
- * Best-effort: a missing tree or a newer pi-ai that already dropped the lines
- * only warns; it never fails the install.
- */
-import { existsSync, readFileSync, writeFileSync } from 'node:fs'
-import { dirname, resolve } from 'node:path'
-import { createRequire } from 'node:module'
-import { fileURLToPath } from 'node:url'
-
-const require = createRequire(import.meta.url)
-
-/** Locate the installed pi-ai root without crossing its `exports` map. */
-function findPiRoot() {
-  try {
-    const main = require.resolve('@earendil-works/pi-ai')
-    if (basenameIsDist(dirname(main))) return dirname(dirname(main))
-  } catch {
-    // Fall through to the layout probe below.
-  }
-  const probe = resolve(dirname(fileURLToPath(import.meta.url)), '..', 'node_modules', '@earendil-works', 'pi-ai')
-  return existsSync(`${probe}/package.json`) ? probe : undefined
-}
-
-/** True when `dir` is the package's `dist` directory. */
-function basenameIsDist(dir) {
-  return dir.endsWith('/dist') || dir.endsWith('\\dist')
-}
-
-/**
- * One hunk: `before` (kept) immediately followed by `removed`.
- * Whitespace between them is flexible; nothing else is touched.
- */
-const HUNKS = [
-  ['dist/api/anthropic-messages.js',
-    'block.partialJson += event.delta.partial_json;',
-    'block.arguments = parseStreamingJson(block.partialJson);'],
-  ['dist/api/bedrock-converse-stream.js',
-    'block.partialJson = (block.partialJson || "") + (delta.toolUse.input || "");',
-    'block.arguments = parseStreamingJson(block.partialJson);'],
-  ['dist/api/mistral-conversations.js',
-    'block.partialArgs = (block.partialArgs || "") + argsDelta;',
-    'block.arguments = parseStreamingJson(block.partialArgs);'],
-  ['dist/api/openai-completions.js',
-    'block.partialArgs = (block.partialArgs ?? "") + toolCall.function.arguments;',
-    'block.arguments = parseStreamingJson(block.partialArgs);'],
-  ['dist/api/openai-responses-shared.js',
-    'slot.block.partialJson += event.delta;',
-    'slot.block.arguments = parseStreamingJson(slot.block.partialJson);'],
-]
-
-const escapeRegExp = text => text.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
-
-const piRoot = findPiRoot()
-if (piRoot === undefined) {
-  console.warn('[dshopencodego] patch-pi-ai: pi-ai not installed, skipping')
-  process.exit(0)
-}
-
-let version = 'unknown'
-try {
-  version = JSON.parse(readFileSync(`${piRoot}/package.json`, 'utf8')).version ?? 'unknown'
-} catch (error) {
-  console.warn(`[dshopencodego] patch-pi-ai: cannot read pi-ai version, skipping (${String(error)})`)
-  process.exit(0)
-}
-if (version !== '0.87.1') {
-  console.warn(`[dshopencodego] patch-pi-ai: pi-ai ${version} is not 0.87.1, skipping (re-check upstream pi#9265)`)
-  process.exit(0)
-}
-
-let patched = 0
-let files = 0
-for (const [rel, before, removed] of HUNKS) {
-  const path = `${piRoot}/${rel}`
-  if (!existsSync(path)) continue
-  files += 1
-  const pattern = new RegExp(`(${escapeRegExp(before)}\\n)(\\s*${escapeRegExp(removed)}\\n)`)
-  const text = readFileSync(path, 'utf8')
-  if (!pattern.test(text)) continue
-  writeFileSync(path, text.replace(pattern, '$1'))
-  patched += 1
-}
-
-// pi-messages.js spans the delta assignment over two lines.
-{
-  const path = `${piRoot}/dist/api/pi-messages.js`
-  if (existsSync(path)) {
-    files += 1
-    const pattern = /partial\.content\[event\.contentIndex\]\.arguments =\s*\n?\s*parseStreamingJson\(json\);\n/
-    const text = readFileSync(path, 'utf8')
-    if (pattern.test(text)) {
-      writeFileSync(path, text.replace(pattern, ''))
-      patched += 1
-    }
-  }
-}
-console.log(`[dshopencodego] patch-pi-ai: ${patched} hunks applied across ${files} files (idempotent)`)