mirror of
http://192.168.0.88:13333/lywsvip/openclaw-zero-token.git
synced 2026-06-01 22:51:24 +08:00
Web model fixes: - ChatGPT: update send button selectors for 2026 UI - Gemini: use innerText instead of textContent, tighten response selectors - Kimi: fix Connect binary protocol parser (only accept append/set ops) - Qwen CN: add credentials:"include" to fix 403 signature error - Perplexity: rewrite client with Playwright native keyboard API, only send last user message instead of full system prompt - DeepSeek: fix tool_call regex, add data.v array handling, XML argument fallback, JUNK_TOKENS Unicode variants E2E test suite (scripts/): - Two-phase runner: TUI first (L2 HTTP + L3 WebSocket), then WebUI (L5 Playwright) - Anti-ban: adaptive delay with provider interleaving - Message pool: 20 natural CN/EN messages, random selection - HTML report output to reports/ Other: - Simplify chat UI: remove session/agent selector, keep only model selector - onboard webauth: auto-exit after auth completes (process.exit) - Move vitest config to scripts/ directory Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
21 lines
625 B
Bash
21 lines
625 B
Bash
# shellcheck shell=bash
|
||
# 由 test-web-model-matrix.sh 和 test-web-model-e2e.sh source;
|
||
# 定义待测 provider/model(与 src/zero-token/bridge/web-providers.ts 目录一致)。
|
||
|
||
# 完整 Web 模型矩阵(13 个 provider)
|
||
WEB_MODEL_MATRIX_ENTRIES=(
|
||
"claude-web/claude-sonnet-4-6"
|
||
"chatgpt-web/gpt-4"
|
||
"deepseek-web/deepseek-chat"
|
||
"doubao-web/doubao-seed-2.0"
|
||
"qwen-web/qwen3.5-plus"
|
||
"qwen-cn-web/Qwen3.5-Plus"
|
||
"kimi-web/moonshot-v1-32k"
|
||
"gemini-web/gemini-pro"
|
||
"grok-web/grok-2"
|
||
"glm-web/glm-4-plus"
|
||
"glm-intl-web/glm-4-plus"
|
||
"perplexity-web/perplexity-web"
|
||
"xiaomimo-web/xiaomimo-chat"
|
||
)
|