MacOS下VSCode任务中使用conda环境报错的解决求助

配置了如下tasks.json文件：

{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "conda info",
      "type": "shell",
      "command": "type conda; which conda",
    },
    {
      "label": "Activate conda envrionment: mkdocs",
      "type": "shell",
      "command": "conda",
      "args": [
        "activate", "mkdocs"
      ]
    },
    {
      "label": "Serve mkdocs",
      "type": "shell",
      "command": "mkdocs serve --strict --dev-addr localhost:8010",
    },
    {
      "label": "run mkdocs serve in environment",
      "group": "test",
      "presentation": {
        "reveal": "always",
        "panel": "shared"
      },
      "runOptions": { 
        "runOn": "folderOpen" 
      },
      "dependsOrder": "sequence",
      "dependsOn": [
        "conda info",
        "Activate conda envrionment: mkdocs", 
        "Serve mkdocs"
      ],
    }
  ]
}

运行任务时总是报错：

CommandNotFoundError: Your shell has not been properly configured to use 'conda activate'.

已设置"python.condaPath"，conda info任务显示路径正确，且VSCode终端中可正常执行相关命令，以下是针对MacOS系统的解决方法：

解决方法 #

方法1：使用conda run直接在指定环境执行命令（推荐） #

Conda提供的conda run命令可直接在目标环境中运行指定命令，无需单独执行conda activate，避免shell配置问题。修改后的tasks.json如下：

{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "conda info",
      "type": "shell",
      "command": "type conda; which conda",
    },
    {
      "label": "Serve mkdocs in conda env",
      "type": "shell",
      "command": "conda",
      "args": [
        "run",
        "--name", "mkdocs",
        "mkdocs", "serve", "--strict", "--dev-addr", "localhost:8010"
      ]
    },
    {
      "label": "run mkdocs serve in environment",
      "group": "test",
      "presentation": {
        "reveal": "always",
        "panel": "shared"
      },
      "runOptions": { 
        "runOn": "folderOpen" 
      },
      "dependsOrder": "sequence",
      "dependsOn": [
        "conda info",
        "Serve mkdocs in conda env"
      ],
    }
  ]
}

方法2：加载Conda初始化脚本并合并任务 #

VSCode任务的shell是非交互式的，不会自动加载conda配置，可先手动加载初始化脚本，再执行激活和运行命令，需将激活与运行合并为同一个任务（每个任务是独立shell进程，前一个任务的环境不会传递给后续任务）：

{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "conda info",
      "type": "shell",
      "command": "type conda; which conda",
    },
    {
      "label": "Serve mkdocs",
      "type": "shell",
      // 若使用Anaconda，路径改为$HOME/anaconda3/etc/profile.d/conda.sh
      "command": "source $HOME/miniconda3/etc/profile.d/conda.sh && conda activate mkdocs && mkdocs serve --strict --dev-addr localhost:8010"
    },
    {
      "label": "run mkdocs serve in environment",
      "group": "test",
      "presentation": {
        "reveal": "always",
        "panel": "shared"
      },
      "runOptions": { 
        "runOn": "folderOpen" 
      },
      "dependsOrder": "sequence",
      "dependsOn": [
        "conda info",
        "Serve mkdocs"
      ],
    }
  ]
}

方法3：配置交互式shell #

让任务使用交互式shell，自动加载.zshrc或.bashrc中的conda配置，同样需要合并激活与运行命令：

{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "conda info",
      "type": "shell",
      "command": "type conda; which conda",
    },
    {
      "label": "Serve mkdocs",
      "type": "shell",
      "command": "conda activate mkdocs && mkdocs serve --strict --dev-addr localhost:8010",
      "options": {
        "shell": {
          "executable": "/bin/zsh",
          "args": ["-i", "-c"]
        }
      }
    },
    {
      "label": "run mkdocs serve in environment",
      "group": "test",
      "presentation": {
        "reveal": "always",
        "panel": "shared"
      },
      "runOptions": { 
        "runOn": "folderOpen" 
      },
      "dependsOrder": "sequence",
      "dependsOn": [
        "conda info",
        "Serve mkdocs"
      ],
    }
  ]
}

内容的提问来源于stack exchange，提问作者Stefan