Add `--no-trunc` option to `docker stack deploy` for full error messages

从 docker stack deploy 的 CLI 入口点开始，将其输出处理与现有的 docker service ps --no-trunc 实现进行比较。跟踪部署状态和错误输出，然后验证新选项是否保留完整消息，以及相关的 CLI 测试是否通过。

Description

Problem

When running docker stack deploy, deployment errors (especially scheduling constraint failures) are truncated in the CLI output. For example:

1/1: no suitable node (scheduling constraints not satisfied on 1 node; host-mod…
overall progress: 0 out of 1 tasks

This truncation makes it difficult to understand why a service could not be scheduled, as the most relevant part of the message is often cut off. This behavior appears to be hardcoded and does not respect terminal width or environment variables like COLUMNS.

Workarounds

• Manually running docker service ps <service> --no-trunc after deployment
• Inspecting tasks individually with docker inspect
• Writing custom scripts to combine deployment and diagnostic commands

These add friction and delay in understanding deployment issues.

Proposal

Add a --no-trunc flag to docker stack deploy, similar to the one already available in docker service ps, which disables truncation of deployment status and error messages.

Other alternatives could include:

• Respecting the actual terminal width or COLUMNS environment variable
• Automatically disabling truncation when output is non-interactive (e.g. in CI pipelines)
• Supporting full output when --verbose or --debug is enabled

Benefits

• Immediate visibility into deployment failures without requiring follow-up commands
• Better support for CI/CD pipelines where error visibility is critical
• Makes docker stack deploy consistent with other CLI commands that support --no-trunc

Related

• docker service ps already supports --no-trunc and behaves as expected
• No current CLI flags exist to control truncation during stack deployment