Skip to content

fix(codexmate): harden bridge retries and OpenClaw UX #207

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ymkiux merged 13 commits into from
Jul 13, 2026
+872 −295
Merged

fix(codexmate): harden bridge retries and OpenClaw UX #207

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
7628878
fix(openai-bridge): avoid duplicate error headers
awsl233777 Jul 12, 2026
25d4f27
fix(openai-bridge): close failed post-header responses
awsl233777 Jul 12, 2026
563d3bf
fix(openai-bridge): retry transient HTTP statuses
awsl233777 Jul 12, 2026
4372218
fix(openai-bridge): retry transient responses streams
awsl233777 Jul 12, 2026
6431495
fix(web-ui): expand main area when sidebar collapses
ymkiux Jul 12, 2026
187a747
fix(openai-bridge): hide retry counts and keep retrying
awsl233777 Jul 12, 2026
5c9c66d
fix(openai-bridge): align transient retries with maxx
awsl233777 Jul 12, 2026
f1429fa
feat(openclaw): enrich config tab diagnostics
awsl233777 Jul 13, 2026
3c684cb
style(openclaw): tighten workspace editor copy
awsl233777 Jul 13, 2026
c8496d3
style(openclaw): align workspace file cards
awsl233777 Jul 13, 2026
c34d962
style(openclaw): remove redundant workspace input
awsl233777 Jul 13, 2026
de46497
style(openclaw): simplify workspace card labels
awsl233777 Jul 13, 2026
aec7088
chore(release): bump version to 0.1.2
awsl233777 Jul 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 17 additions & 9 deletions cli/openai-bridge-retry.js
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,14 +1,20 @@
const DEFAULT_BRIDGE_MAX_RETRIES = 2;
const MIN_BRIDGE_MAX_RETRIES = 2;
const MAX_BRIDGE_MAX_RETRIES = 10;
const DEFAULT_BRIDGE_MAX_RETRIES = Infinity;
const BASE_TRANSIENT_RETRY_DELAY_MS = 200;
const MAX_TRANSIENT_RETRY_DELAY_MS = 5000;

function normalizeBridgeMaxRetries(value, fallback = DEFAULT_BRIDGE_MAX_RETRIES) {
const raw = Number(value);
const fallbackRaw = Number(fallback);
const base = Number.isFinite(raw) ? raw : (Number.isFinite(fallbackRaw) ? fallbackRaw : DEFAULT_BRIDGE_MAX_RETRIES);
return Math.min(MAX_BRIDGE_MAX_RETRIES, Math.max(MIN_BRIDGE_MAX_RETRIES, Math.floor(base)));
if (Number.isFinite(raw) && raw >= 0) return Math.floor(raw);
if (Number.isFinite(fallbackRaw) && fallbackRaw >= 0) return Math.floor(fallbackRaw);
return DEFAULT_BRIDGE_MAX_RETRIES;
}


function isTransientHttpStatus(status) {
const code = Number(status);
if (code === 408 || code === 409 || code === 425 || code === 429) return true;
return code >= 500 && code <= 599;
}

function isTransientNetworkError(error) {
Expand All @@ -30,7 +36,7 @@ function getTransientRetryDelayMs(attempt) {
async function retryTransientRequest(executor, options = {}) {
const maxRetries = normalizeBridgeMaxRetries(options && options.maxRetries);
let lastResult = null;
for (let attempt = 0; attempt <= maxRetries; attempt += 1) {
for (let attempt = 0; attempt === 0 || attempt <= maxRetries; attempt += 1) {
if (attempt > 0) {
const delay = getTransientRetryDelayMs(attempt);
// eslint-disable-next-line no-await-in-loop
Expand All @@ -43,20 +49,22 @@ async function retryTransientRequest(executor, options = {}) {
const result = await executor(attempt);
lastResult = result;
if (!result) return result;
if (result.status && result.status > 0) {
if (!isTransientHttpStatus(result.status)) return result;
continue;
}
if (result.ok) return result;
if (result.retry) return result;
if (result.status && result.status > 0) return result;
if (!isTransientNetworkError(result.error)) return result;
}
return lastResult;
}

module.exports = {
DEFAULT_BRIDGE_MAX_RETRIES,
MIN_BRIDGE_MAX_RETRIES,
MAX_BRIDGE_MAX_RETRIES,
normalizeBridgeMaxRetries,
isTransientNetworkError,
isTransientHttpStatus,
getTransientRetryDelayMs,
retryTransientRequest
};
15 changes: 14 additions & 1 deletion cli/openai-bridge-runtime.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1472,7 +1472,20 @@ function streamChatCompletionsAsResponsesSse(targetUrl, options = {}) {
response: {
id: state.responseId,
model: state.model,
created_at: state.createdAt
created_at: state.createdAt,
status: 'in_progress',
output: []
}
});
writeSse(res, 'response.in_progress', {
type: 'response.in_progress',
sequence_number: state.nextSeq(),
response: {
id: state.responseId,
model: state.model,
created_at: state.createdAt,
status: 'in_progress',
output: []
Comment on lines +1475 to +1488

@coderabbitai coderabbitai Bot Jul 12, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🎯 Functional Correctness | 🟠 Major | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
set -eu
sed -n '1468,1492p' cli/openai-bridge-runtime.js
rg -n -A8 -B4 'response\.created|response\.in_progress|sequence_number' tests/unit/openai-bridge-upstream-responses.test.mjs

Repository: SakuraByteCore/codexmate

Length of output: 997

🏁 Script executed:

#!/bin/bash
set -eu

printf '\n== response event emitters ==\n'
rg -n -A6 -B4 "writeSse\\(res, 'response\\.(created|in_progress|completed|failed|incomplete)'" cli/openai-bridge-runtime.js

printf '\n== sequence number helpers/usages ==\n'
rg -n "nextSeq\\(|sequence_number" cli tests --glob '!**/node_modules/**'

printf '\n== response streaming tests ==\n'
sed -n '1,260p' tests/unit/openai-bridge-upstream-responses.test.mjs

Repository: SakuraByteCore/codexmate

Length of output: 17661

🏁 Script executed:

#!/bin/bash
set -eu

printf '\n== shared Responses SSE emitter ==\n'
sed -n '800,920p' cli/openai-bridge-runtime.js

printf '\n== tests mentioning response.created or sequence_number ==\n'
rg -n -A4 -B4 "response\\.created|response\\.in_progress|sequence_number" tests cli/openai-bridge-runtime.js cli/openai-bridge.js

printf '\n== response SSE parsing helpers in tests ==\n'
rg -n -A5 -B5 "event: 'response\\.|type: 'response\\.|sequence_number" tests/unit/openai-bridge-upstream-responses.test.mjs

Repository: SakuraByteCore/codexmate

Length of output: 18992

🌐 Web query:

OpenAI Responses streaming schema response.created sequence_number required response.in_progress

💡 Result:

In the OpenAI Responses API, streaming events follow a structured schema that includes both type and sequence_number fields [1][2]. Key Event Types: - response.created: Emitted immediately after a response is created. It includes the response object and the sequence_number [2][3]. - response.in_progress: Emitted while the response is being generated. Like other lifecycle events, it includes the response object and the sequence_number [2][3]. Sequence Number: The sequence_number is a required integer field for streaming events in the Responses API [1][4][5]. It is monotonically increasing for each stream and is used by clients to enforce correct event ordering and to ignore duplicate events [6]. Documentation and Implementation: Official OpenAI documentation defines these events with a specific schema [1]. Client libraries, such as the official Node.js SDK, define these events in their type definitions (e.g., ResponseCreatedEvent and ResponseInProgressEvent) as requiring both a type and a sequence_number [7][2]. While the API specification mandates these fields, developers have noted that implementation behavior can occasionally deviate from the spec in specific scenarios (such as error events), requiring robust handling in client implementations [5]. Note: This structure specifically applies to the Responses API (often used in the context of the Assistants API/Realtime capabilities) and differs from the traditional Chat Completions streaming API, which utilizes a different chunk-based structure (chat.completion.chunk) [8][9][10].

Citations:

🏁 Script executed:

#!/bin/bash
set -eu

printf '\n== local docs for Responses sequencing ==\n'
rg -n "sequence_number|response\\.created|response\\.in_progress|Responses streaming|streaming schema" README* docs cli tests --glob '!**/node_modules/**'

printf '\n== all response.created emitters ==\n'
rg -n -A5 -B3 "writeSse\\(res, 'response\\.created'" cli tests

Repository: SakuraByteCore/codexmate

Length of output: 4152

Add sequence_number to response.created. This lifecycle event is missing the top-level ordering field, so it doesn’t match the Responses streaming schema; response.in_progress should keep the next sequence value.

Proposed fix 
                     created_at: state.createdAt,
                     status: 'in_progress',
                     output: []
-                }
+                },
+                sequence_number: state.nextSeq()
             });
📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change
created_at: state.createdAt,
status: 'in_progress',
output: []
}
});
writeSse(res, 'response.in_progress', {
type: 'response.in_progress',
sequence_number: state.nextSeq(),
response: {
id: state.responseId,
model: state.model,
created_at: state.createdAt,
status: 'in_progress',
output: []
created_at: state.createdAt,
status: 'in_progress',
output: []
},
sequence_number: state.nextSeq()
});
writeSse(res, 'response.in_progress', {
type: 'response.in_progress',
sequence_number: state.nextSeq(),
response: {
id: state.responseId,
model: state.model,
created_at: state.createdAt,
status: 'in_progress',
output: []
🤖 Prompt for AI Agents 
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@cli/openai-bridge-runtime.js` around lines 1475 - 1488, Add the next sequence
value to the top-level payload of the response.created event, matching the
Responses streaming schema. Keep response.in_progress using its existing
state.nextSeq() call and do not alter its ordering behavior.

Source: MCP tools

}
});

Expand Down
30 changes: 20 additions & 10 deletions cli/openai-bridge.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,7 +87,6 @@ function normalizeOpenaiUpstreamBaseUrl(rawValue) {
}

const {
normalizeBridgeMaxRetries,
parseJsonOrError,
extractChatCompletionResult,
convertResponsesRequestToChatCompletions,
Expand All @@ -114,11 +113,10 @@ function normalizeUpstreamEntry(entry) {
const apiKey = normalizeText(entry.apiKey || entry.api_key || entry.key || '');
const headersRaw = entry.headers || entry.extraHeaders || entry.extra_headers || null;
const headers = normalizeHeadersMap(headersRaw);
const maxRetries = normalizeBridgeMaxRetries(entry.maxRetries ?? entry.max_retries);
if (!baseUrl || !isValidHttpUrl(baseUrl)) {
return null;
}
return { baseUrl, apiKey, headers, maxRetries };
return { baseUrl, apiKey, headers };
}

function normalizeHeadersMap(value) {
Expand Down Expand Up @@ -169,7 +167,6 @@ function upsertOpenaiBridgeProvider(filePath, providerName, upstreamBaseUrl, api
const baseUrl = normalizeOpenaiUpstreamBaseUrl(upstreamBaseUrl);
const key = normalizeText(apiKey);
const nextHeaders = normalizeHeadersMap(headers);
const maxRetries = normalizeBridgeMaxRetries(options && options.maxRetries);

if (!name) {
return { error: 'Provider name is required' };
Expand All @@ -191,7 +188,6 @@ function upsertOpenaiBridgeProvider(filePath, providerName, upstreamBaseUrl, api
baseUrl,
apiKey: key,
headers: Object.keys(nextHeaders).length ? nextHeaders : existingHeaders,
maxRetries
}
}
};
Expand Down Expand Up @@ -384,7 +380,7 @@ function createOpenaiBridgeHttpHandler(options = {}) {
maxBytes: maxUpstreamBytes,
httpAgent,
httpsAgent
}), { maxRetries: upstream.maxRetries });
}));
if (!result.ok) {
res.writeHead(502, { 'Content-Type': 'application/json; charset=utf-8' });
res.end(JSON.stringify({ error: `Upstream request failed: ${result.error}` }));
Expand Down Expand Up @@ -446,7 +442,7 @@ function createOpenaiBridgeHttpHandler(options = {}) {
res,
model: typeof chatBody.model === 'string' ? chatBody.model : '',
toolTypesByName: converted.toolTypesByName || {}
}), { maxRetries: upstream.maxRetries });
}));
if (!streamed.ok) {
if (res.writableEnded || res.destroyed) {
return;
Expand All @@ -471,7 +467,7 @@ function createOpenaiBridgeHttpHandler(options = {}) {
maxBytes: maxUpstreamBytes,
httpAgent,
httpsAgent
}), { maxRetries: upstream.maxRetries });
}));
if (!upstreamResult.ok) {
res.writeHead(502, { 'Content-Type': 'application/json; charset=utf-8' });
res.end(JSON.stringify({ error: `Upstream request failed: ${upstreamResult.error}` }));
Expand Down Expand Up @@ -514,8 +510,23 @@ function createOpenaiBridgeHttpHandler(options = {}) {
res.writeHead(200, { 'Content-Type': 'application/json; charset=utf-8' });
res.end(JSON.stringify(ensureResponseMetadata(responsesPayload)));
} catch (e) {
if (res.writableEnded || res.destroyed) {
return;
}
const message = e && e.message ? e.message : 'Internal Error';
if (res.headersSent) {
try {
res.end();
} catch (_) {
// Headers are already committed. Close the socket instead of leaving the client waiting forever.
if (!res.destroyed && typeof res.destroy === 'function') {
res.destroy(e);
}
}
return;
}
res.writeHead(500, { 'Content-Type': 'application/json; charset=utf-8' });
res.end(JSON.stringify({ error: e && e.message ? e.message : 'Internal Error' }));
res.end(JSON.stringify({ error: message }));
}
})();

Expand Down Expand Up @@ -543,7 +554,6 @@ module.exports = {
extractChatCompletionResult,
buildResponsesPayloadFromChatResult,
retryTransientRequest,
normalizeBridgeMaxRetries,
normalizeOpenaiUpstreamBaseUrl,
extractResponsesOutputText,
shouldFallbackFromUpstreamResponses,
Expand Down
4 changes: 2 additions & 2 deletions package-lock.json
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

2 changes: 1 addition & 1 deletion package.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
{
"name": "codexmate",
"version": "0.1.1",
"version": "0.1.2",
"description": "Codex/Claude Code/OpenClaw 配置、会话与任务编排 CLI + Web 工具",
"main": "cli.js",
"bin": {
Expand Down
Loading
Loading