Continue appears to rewrite a configured localhost API base URL to 127.0.0.1. (WSL)
Los mantenedores suelen responder en 8 días
Nadie ha tomado este issue todavía.
Evaluación
- Dificultad
- 3/5
- Tiempo estimado
- 1-2 días
- Aptitud para principiantes
- 65/100
- Tipo de issue
- Error
- Claridad
- Bien especificado
- Estado de actividad
- Activo
- Stack tecnológico
- typescript
- Área
- backend, networking
Línea de trabajo
The report shows Continue requesting http://127.0.0.1:8080/completion even though apiBase is configured as http://localhost:8080, while http://[::1]:8080 works. Reproduce with the llama.cpp provider config given, then find where an apiBase URL is normalized, since the config is being rewritten before the request is built. Done looks like the user-supplied hostname being preserved so the request goes to localhost:8080/completion.
Escrito por el modelo de indexación a partir del texto del issue.
Descripción
Before submitting your bug report
- I've tried finding an answer on the [Continue docs site](https://docs.continue.dev/)
- I'm not able to find an [open issue](https://github.com/continuedev/continue/issues?q=is%3Aopen+is%3Aissue) that reports the same bug
- I've seen the [troubleshooting guide](https://docs.continue.dev/troubleshooting) on the Continue Docs
Relevant environment info
OS: Windows with WSL2 (Ubuntu)
Continue version: V2.0.0
IDE version: VS Code 1.140.0
Model: Qwen3 14B, llama.cpp, ROCm
config:
models:
- name: Qwen3 14B (llama-rocm)
provider: llama.cpp
model: /models/Qwen3-14B-Q4_K_M.gguf
apiBase: http://localhost:8080
roles:
- chat
- edit
- apply
defaultCompletionOptions:
temperature: 0.2
maxTokens: 4096
Description
Continue appears to replace an explicitly configured localhost API base with 127.0.0.1.
My llama.cpp server is reachable through localhost and IPv6 loopback, but not through IPv4 loopback:
curl http://localhost:8080/health
# works
curl http://[::1]:8080/health
# works
curl http://127.0.0.1:8080/health
# fails
Continue is explicitly configured with:
apiBase: http://localhost:8080
However, when Continue attempts a completion, the error shows a request being made to:
http://127.0.0.1:8080/completion
This causes the request to fail with ECONNREFUSED.
The issue appears to be that Continue is replacing localhost with 127.0.0.1 instead of preserving the user-provided hostname and allowing the operating system or runtime to resolve it.
Using IPv6 explicitly works around the issue:
apiBase: http://[::1]:8080
With that configuration, Continue can connect to the llama.cpp server successfully.
Expected behavior
If apiBase is explicitly configured as:
http://localhost:8080
Continue should preserve that hostname and connect to:
http://localhost:8080/completion
It should not rewrite localhost to 127.0.0.1.
localhost and 127.0.0.1 are not always interchangeable, particularly in environments involving WSL2, IPv6, containers, and virtualized networking.
To reproduce
-
Run a llama.cpp server that is reachable through
localhostand::1, but not through127.0.0.1. -
Verify the endpoints:
curl http://localhost:8080/health
curl http://[::1]:8080/health
curl http://127.0.0.1:8080/health
- Configure Continue with:
models:
- name: Qwen3 14B (llama-rocm)
provider: llama.cpp
model: /models/Qwen3-14B-Q4_K_M.gguf
apiBase: http://localhost:8080
roles:
- chat
- edit
- apply
-
Select the model in Continue and send a chat request.
-
Observe that Continue attempts to connect to:
http://127.0.0.1:8080/completion
rather than:
http://localhost:8080/completion
- Change the configuration to:
apiBase: http://[::1]:8080
- Retry the request and observe that it succeeds.
Log output
request to http://127.0.0.1:8080/completion failed, reason: connect ECONNREFUSED 127.0.0.1:8080
- Lenguaje dominante
- TypeScript
- Estrellas
- 36.5k
- Forks
- 5.6k
- Métricas de merge de PR
- Sin PR fusionados en 30 d
Preparar el entorno
- Sin Dockerfile ni archivo de Docker Compose
- Tiene una plantilla de pull request
- Leer la guía de contribución
Primeros pasos
- Lee el issue completo y luego la guía de contribución del proyecto.
- Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
- Haz un fork del repositorio y trabaja en una rama.
- Abre un pull request que haga referencia al número del issue.
Más de continuedev/continue
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 67/100
continuedev/continue#13486 ·
Los mantenedores suelen responder en 8 días
-
Chat panel overflows horizontally when VS Code zoom level is increased (Cmd/Ctrl + +)Posiblemente ocupada @tang-ontheway la tomó hace 1 día. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 80/100
continuedev/continue#13483 ·
Los mantenedores suelen responder en 8 días
-
Docs: show tool selection for a shared remote MCP endpointPosiblemente ocupada Un pull request vinculado a esta issue está abierto o ya se fusionó. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 84/100
continuedev/continue#13464 · 1 comentario ·
Los mantenedores suelen responder en 8 días
-
Inline code becomes unreadable in VS Code high-contrast themesPosiblemente ocupada @aminesayagh la tomó hace 21 días. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 63/100
continuedev/continue#13424 ·
Los mantenedores suelen responder en 8 días
-
create_new_file writes a JSON file on one line when the model sends its contents as a JSON valuePosiblemente ocupada Un pull request vinculado a esta issue está abierto o ya se fusionó. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 72/100
continuedev/continue#13363 ·
Los mantenedores suelen responder en 8 días
Todos los issues de continuedev/continue
Issues similares
-
status/need-triage
Dificultad 2/5 1-3 horas Aptitud para principiantes 76/100
google-gemini/gemini-cli#30328 · 1 comentario ·
Los mantenedores suelen responder en 1 día
-
[hooks] SessionStart injects the knowledge-update frontmatter on a CRLF checkoutPosiblemente ocupada @abhiramshaji-cyber la tomó hoy. Abierto
Dificultad 2/5 1-3 horas Aptitud para principiantes 86/100
vercel/vercel-plugin#429 ·
Los mantenedores suelen responder en 1 día
-
severity:medium triage:confirmed triage:in-progress triage:reproduced
Dificultad 2/5 1-3 horas Aptitud para principiantes 80/100
Cotal-AI/Cotal#7231 · 3 comentarios ·
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 66/100
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 78/100
boxlite-ai/boxlite#2777 ·
Los mantenedores suelen responder en 1 día