[BUG] Error executing vim.schedule lua callback: "src must be a table or function" in `lua/vectorcode/init.lua:194`

[minusfive]

Describe the bug

VectorCode nvim plugin throws errors on start. Seems to continue to work OK, but the errors still show up on every start. After start errors don't appear again, though.

Error executing vim.schedule lua callback: ...Cellar/neovim/0.11.4/share/nvim/runtime/lua/vim/iter.lua:1067: src must be a table or function
stack traceback:
	[C]: in function 'error'
	...Cellar/neovim/0.11.4/share/nvim/runtime/lua/vim/iter.lua:1067: in function 'iter'
	...local/share/nvim/lazy/VectorCode/lua/vectorcode/init.lua:194: in function ''
	vim/_editor.lua: in function <vim/_editor.lua:0>

To Reproduce

Steps to reproduce the behavior:

1. install VectorCode and nvim plugin following Readme
2. vectorcode init in a project
3. open nvim
4. run :VectorCode <anything> in nvim

Expected behavior

No errors

VectorCode Configuration

Note

I don't use a config.json, neither globally nor in the project. I rely on defaults. Not sure if that has anything to do with it, but figured I'd mention it.

For issues with the Neovim plugin, please also attach your setup options:

{
    "Davidyz/VectorCode",
    version = "*", -- optional, depending on whether you're on nightly or release
    build = "uv tool upgrade vectorcode", -- optional but recommended. This keeps your CLI up-to-date.
    dependencies = {
      "nvim-lua/plenary.nvim",
      {
        "neovim/nvim-lspconfig",
        ---@class PluginLspOpts
        opts = {
          servers = { vectorcode_server = {} },
        },
      },
    },
    cmd = { "VectorCode" },
    ---@module "vectorcode"
    ---@type VectorCode.Opts|{}
    opts = {
      async_backend = "lsp",
      on_setup = {
        update = true, -- whether to update the CLI on setup
        lsp = true, -- whether to setup the LSP server
      },
    },
    keys = {
      {
        "<leader>au",
        function() return require("vectorcode").update() end,
        desc = "Update (VectorCode)",
        mode = { "n", "v" },
      },
    },
  }

If it only occurs when you use VectorCode with a particular plugin, please
attach the relevant config here:

Platform information:

• If the issue is about the neovim plugin, attach the neovim version you're using:

0.11.4

System Information:

For Mac users, please also mention whether you're using intel or apple silicon devices.

• OS: MacOS 15.6.1, Apple Silicon (M3)

Additional context
Add any other context about the problem here. Please attach
CLI logs
or
nvim plugin logs
if applicable.

[INFO  Mon Sep 22 09:57:54 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:123: Received setup opts:
 {
  async_backend = "lsp",
  on_setup = {
    lsp = true,
    update = true
  }
}
[DEBUG Mon Sep 22 09:57:54 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/jobrunner/cmd.lua:14: cmd jobrunner for buffer 0 args: { "check", "config" }
[DEBUG Mon Sep 22 09:57:54 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:51: Using vim.lsp.config.vectorcode_server for LSP config:
 {
  capabilities = {
    textDocument = {
      completion = {
        completionItem = {
          commitCharactersSupport = false,
          deprecatedSupport = true,
          documentationFormat = { "markdown", "plaintext" },
          insertReplaceSupport = true,
          insertTextModeSupport = {
            valueSet = { 1 }
          },
          labelDetailsSupport = true,
          preselectSupport = false,
          resolveSupport = {
            properties = { "documentation", "detail", "additionalTextEdits", "command", "data" }
          },
          snippetSupport = true,
          tagSupport = {
            valueSet = { 1 }
          }
        },
        completionList = {
          itemDefaults = { "commitCharacters", "editRange", "insertTextFormat", "insertTextMode", "data" }
        },
        contextSupport = true,
        insertTextMode = 1
      }
    },
    workspace = {
      fileOperations = {
        didRename = true,
        willRename = true
      }
    }
  },
  cmd = { "vectorcode-server" },
  name = "vectorcode_server",
  root_dir = "/Users/<user>/dev/personal/dot",
  root_markers = { ".vectorcode", ".git" },
  settings = {}
}
[INFO  Mon Sep 22 09:57:54 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:143: Finished processing opts:
 {
  async_backend = "lsp",
  async_opts = {
    debounce = 10,
    events = { "BufWritePost", "InsertEnter", "BufReadPost" },
    exclude_this = true,
    n_query = 1,
    notify = true,
    query_cb = <function 1>,
    run_on_register = false,
    single_job = false,
    timeout_ms = 5000
  },
  cli_cmds = {
    vectorcode = "vectorcode"
  },
  exclude_this = true,
  n_query = 1,
  notify = true,
  on_setup = {
    lsp = true,
    update = true
  },
  sync_log_env_var = false,
  timeout_ms = 5000
}
[DEBUG Mon Sep 22 09:57:54 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:155: Using LSP backend for cacher.

[minusfive]

@Davidyz thanks for the quick turnaround!

Unfortunately it seems the issue persists, though it's moved:

Error executing vim.schedule lua callback: ...local/share/nvim/lazy/VectorCode/lua/vectorcode/init.lua:154: attempt to index a nil value
stack traceback:
	...local/share/nvim/lazy/VectorCode/lua/vectorcode/init.lua:154: in function 'update'
	...cal/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:99: in function 'stdout_cb'
	...local/share/nvim/lazy/VectorCode/lua/vectorcode/init.lua:194: in function ''
	vim/_editor.lua: in function <vim/_editor.lua:0>

Here's the full log:

[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:123: Received setup opts:
 {
  async_backend = "lsp",
  on_setup = {
    lsp = true,
    update = true
  }
}
[DEBUG Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/jobrunner/cmd.lua:14: cmd jobrunner for buffer 0 args: { "check", "config" }
[DEBUG Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:51: Using vim.lsp.config.vectorcode_server for LSP config:
 {
  capabilities = {
    textDocument = {
      completion = {
        completionItem = {
          commitCharactersSupport = false,
          deprecatedSupport = true,
          documentationFormat = { "markdown", "plaintext" },
          insertReplaceSupport = true,
          insertTextModeSupport = {
            valueSet = { 1 }
          },
          labelDetailsSupport = true,
          preselectSupport = false,
          resolveSupport = {
            properties = { "documentation", "detail", "additionalTextEdits", "command", "data" }
          },
          snippetSupport = true,
          tagSupport = {
            valueSet = { 1 }
          }
        },
        completionList = {
          itemDefaults = { "commitCharacters", "editRange", "insertTextFormat", "insertTextMode", "data" }
        },
        contextSupport = true,
        insertTextMode = 1
      }
    },
    workspace = {
      fileOperations = {
        didRename = true,
        willRename = true
      }
    }
  },
  cmd = { "vectorcode-server" },
  name = "vectorcode_server",
  root_dir = "/Users/<user>/dev/personal/dot",
  root_markers = { ".vectorcode", ".git" },
  settings = {}
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:143: Finished processing opts:
 {
  async_backend = "lsp",
  async_opts = {
    debounce = 10,
    events = { "BufWritePost", "InsertEnter", "BufReadPost" },
    exclude_this = true,
    n_query = 1,
    notify = true,
    query_cb = <function 1>,
    run_on_register = false,
    single_job = false,
    timeout_ms = 5000
  },
  cli_cmds = {
    vectorcode = "vectorcode"
  },
  exclude_this = true,
  n_query = 1,
  notify = true,
  on_setup = {
    lsp = true,
    update = true
  },
  sync_log_env_var = false,
  timeout_ms = 5000
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:63: Received codecompanion extension opts:
 {
  add_slash_command = true,
  add_tool = true,
  prompt_library = {
    ["Neovim Tutor"] = {
      file_patterns = { "lua/**/*.lua", "doc/**/*.txt" },
      project_root = "/opt/homebrew/Cellar/neovim/0.11.4/share/nvim/runtime"
    }
  },
  tool_group = {
    collapse = true,
    enabled = true,
    extras = { "file_search" }
  },
  tool_opts = {
    files_ls = {},
    files_rm = {},
    ls = {
      include_in_toolbox = true
    },
    query = {
      include_in_toolbox = true
    },
    use_lsp = true,
    vectorise = {
      include_in_toolbox = true
    }
  }
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/integrations/codecompanion/ls_tool.lua:18: Loading `vectorcode_ls` with the following opts:
{
  include_in_toolbox = true,
  requires_approval = false,
  use_lsp = true
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/integrations/codecompanion/common.lua:44: Using LSP runner for CodeCompanion tool.
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:90: vectorcode_ls tool has been created.
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/integrations/codecompanion/query_tool.lua:140: Loading `vectorcode_query` with the following opts:
{
  chunk_mode = false,
  default_num = 10,
  include_in_toolbox = true,
  max_num = -1,
  no_duplicate = true,
  requires_approval = false,
  summarise = {
    enabled = false,
    query_augmented = true,
    system_prompt = "You are an expert and experienced code analyzer and summarizer. Your primary task is to analyze provided source code, which will be given as a list of XML objects, and generate a comprehensive, well-structured Markdown summary. This summary will serve as a concise source of information for others to quickly understand how the code works and how to interact with it, without needing to delve into the full source code.\n\nInput Format:\nEach XML object represents either a full file or a chunk of a file, containing the following tags:\n- `<path>...</path>`: The absolute file path of the source code.\n- `<document>...</document>`: The full content of a source code file. This tag will not coexist with `<chunk>`.\n- `<chunk>...</chunk>`: A segment of source code from a file. This tag will not coexist with `<document>`.\n- `<start_line>...</start_line>` and `<end_line>...</end_line>`: These tags will be present only when a `<chunk>` tag is used, indicating the starting and ending line numbers of the chunk within its respective file.\n\nYour goal is to process each of these XML objects. If multiple chunks belong to the same file, you must synthesize them to form a cohesive understanding of that file. Generate a single Markdown summary that combines insights from all provided objects.\n\nMarkdown Structure:\n\n    Top-Level Header (#): The absolute or relative file path of the source code.\n\n    Secondary Headers (##): For each top-level symbol (e.g., functions, classes, global variables) defined directly within the source code file that are importable or includable by other programs.\n\n    Tertiary Headers (###): For symbols nested one level deep within a secondary header's symbol (e.g., methods within a class, inner functions).\n\n    Quaternary Headers (####): For symbols nested two levels deep (e.g., a function defined within a method of a class).\n\n    Continue this pattern, incrementing the header level for each deeper level of nesting.\n\nContent for Each Section:\n\n    Descriptive Summary: Each header section (from secondary headers downwards) must contain a concise and informative summary of the symbol defined by that header.\n\n        For Functions/Methods: Explain their purpose, parameters (including types), return values (including types), high-level implementation details, and any significant side effects or core logic. For example, if summarizing a sorting function, include the sorting algorithm used. If summarizing a function that makes an HTTP request, mention the network library employed.\n\n        For Classes: Describe the class's role, its main responsibilities, and key characteristics.\n\n        For Variables (global or within scope): State their purpose, type (if discernible), and initial value or common usage.\n\n        For Modules/Files (under the top-level header): Provide an overall description of the file's purpose, its main components, and its role within the larger project (if context is available).\n\nGeneral Guidelines:\n\n    Clarity and Conciseness: Summaries should be easy to understand, avoiding jargon where possible, and as brief as possible while retaining essential information. The full summary MUST NOT be longer than the original code input. When quoting a symbol in the code, include the line numbers where possible.\n\n    Accuracy: Ensure the summary accurately reflects the code's functionality.\n\n    Focus on Public Interface/Behavior: Prioritize describing what a function/class does and how it's used. Only include details about symbols (variables, functions, classes) that are importable/includable by other programs. DO NOT include local variables and functions that are not accessible by other functions outside their immediate scope.\n\n    No Code Snippets: Do not include any actual code snippets in the summary. Focus solely on descriptive text. If you need to refer to a specific element for context (e.g., in an error description), describe it and provide line numbers for reference from the source code.\n\n    Syntax/Semantic Errors: If the code contains syntax or semantic errors, describe them clearly within the summary, indicating the nature of the error.\n\n    Language Agnostic: Adapt the summary to the specific programming language of the provided source code (e.g., Python, JavaScript, Java, C++, etc.).\n\n    Handle Edge Cases/Dependencies: If a symbol relies heavily on external dependencies or handles specific edge cases, briefly mention these if they are significant to its overall function.\n\n    Information Source: There will be no extra information available to you. Provide the summary solely based on the provided XML objects.\n\n    Omit meaningless results: For an xml object that contains no meaningful information, you're free to omit it, but please leave a sentence in the summary saying that you did this.\n\n    No extra reply: Your reply should solely consist of the summary. Do not say anything else.\n\n    Merge chunks from the same file: When there are chunks that belong to the same file, merge their content so that they're grouped under the same top level header.\n"
  },
  use_lsp = true
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/integrations/codecompanion/query_tool.lua:374: Creating CodeCompanion tool with the following args:
 {
  chunk_mode = false,
  default_num = 10,
  include_in_toolbox = true,
  max_num = -1,
  no_duplicate = true,
  requires_approval = false,
  summarise = {
    enabled = false,
    query_augmented = true,
    system_prompt = "You are an expert and experienced code analyzer and summarizer. Your primary task is to analyze provided source code, which will be given as a list of XML objects, and generate a comprehensive, well-structured Markdown summary. This summary will serve as a concise source of information for others to quickly understand how the code works and how to interact with it, without needing to delve into the full source code.\n\nInput Format:\nEach XML object represents either a full file or a chunk of a file, containing the following tags:\n- `<path>...</path>`: The absolute file path of the source code.\n- `<document>...</document>`: The full content of a source code file. This tag will not coexist with `<chunk>`.\n- `<chunk>...</chunk>`: A segment of source code from a file. This tag will not coexist with `<document>`.\n- `<start_line>...</start_line>` and `<end_line>...</end_line>`: These tags will be present only when a `<chunk>` tag is used, indicating the starting and ending line numbers of the chunk within its respective file.\n\nYour goal is to process each of these XML objects. If multiple chunks belong to the same file, you must synthesize them to form a cohesive understanding of that file. Generate a single Markdown summary that combines insights from all provided objects.\n\nMarkdown Structure:\n\n    Top-Level Header (#): The absolute or relative file path of the source code.\n\n    Secondary Headers (##): For each top-level symbol (e.g., functions, classes, global variables) defined directly within the source code file that are importable or includable by other programs.\n\n    Tertiary Headers (###): For symbols nested one level deep within a secondary header's symbol (e.g., methods within a class, inner functions).\n\n    Quaternary Headers (####): For symbols nested two levels deep (e.g., a function defined within a method of a class).\n\n    Continue this pattern, incrementing the header level for each deeper level of nesting.\n\nContent for Each Section:\n\n    Descriptive Summary: Each header section (from secondary headers downwards) must contain a concise and informative summary of the symbol defined by that header.\n\n        For Functions/Methods: Explain their purpose, parameters (including types), return values (including types), high-level implementation details, and any significant side effects or core logic. For example, if summarizing a sorting function, include the sorting algorithm used. If summarizing a function that makes an HTTP request, mention the network library employed.\n\n        For Classes: Describe the class's role, its main responsibilities, and key characteristics.\n\n        For Variables (global or within scope): State their purpose, type (if discernible), and initial value or common usage.\n\n        For Modules/Files (under the top-level header): Provide an overall description of the file's purpose, its main components, and its role within the larger project (if context is available).\n\nGeneral Guidelines:\n\n    Clarity and Conciseness: Summaries should be easy to understand, avoiding jargon where possible, and as brief as possible while retaining essential information. The full summary MUST NOT be longer than the original code input. When quoting a symbol in the code, include the line numbers where possible.\n\n    Accuracy: Ensure the summary accurately reflects the code's functionality.\n\n    Focus on Public Interface/Behavior: Prioritize describing what a function/class does and how it's used. Only include details about symbols (variables, functions, classes) that are importable/includable by other programs. DO NOT include local variables and functions that are not accessible by other functions outside their immediate scope.\n\n    No Code Snippets: Do not include any actual code snippets in the summary. Focus solely on descriptive text. If you need to refer to a specific element for context (e.g., in an error description), describe it and provide line numbers for reference from the source code.\n\n    Syntax/Semantic Errors: If the code contains syntax or semantic errors, describe them clearly within the summary, indicating the nature of the error.\n\n    Language Agnostic: Adapt the summary to the specific programming language of the provided source code (e.g., Python, JavaScript, Java, C++, etc.).\n\n    Handle Edge Cases/Dependencies: If a symbol relies heavily on external dependencies or handles specific edge cases, briefly mention these if they are significant to its overall function.\n\n    Information Source: There will be no extra information available to you. Provide the summary solely based on the provided XML objects.\n\n    Omit meaningless results: For an xml object that contains no meaningful information, you're free to omit it, but please leave a sentence in the summary saying that you did this.\n\n    No extra reply: Your reply should solely consist of the summary. Do not say anything else.\n\n    Merge chunks from the same file: When there are chunks that belong to the same file, merge their content so that they're grouped under the same top level header.\n"
  },
  use_lsp = true
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:90: vectorcode_query tool has been created.
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/integrations/codecompanion/vectorise_tool.lua:23: Loading `vectorcode_vectorise` with the following opts:
{
  include_in_toolbox = true,
  requires_approval = true,
  use_lsp = true
}
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:90: vectorcode_vectorise tool has been created.
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:90: vectorcode_files_ls tool has been created.
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:90: vectorcode_files_rm tool has been created.
[INFO  Tue Sep 23 10:11:10 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/codecompanion/_extensions/vectorcode/init.lua:107: Loading the following tools into `vectorcode_toolbox` tool group:
{ "vectorcode_ls", "vectorcode_query", "vectorcode_vectorise", "file_search" }
[INFO  Tue Sep 23 10:11:11 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/config.lua:98: Running `vectorcode update` on start up.
[INFO  Tue Sep 23 10:11:11 2025] /Users/<user>/.local/share/nvim/lazy/VectorCode/lua/vectorcode/init.lua:150: vectorcode.update:  Expected value but found invalid token at character 1

[Davidyz]

@minusfive sorry about that. I believe #293 fixes this error. Please give it a try and lmk how it goes.

[minusfive]

@minusfive sorry about that. I believe #293 fixes this error. Please give it a try and lmk how it goes.

Are you kidding? No need to apologize, man. Really appreciate your work!

I'll test and report back.

[minusfive]

@Davidyz commented on the PR. From my perspective this is resolved. Thanks again!

[Bakrog]

Hi @Davidyz, I'm having the same issue when the following is executed during LSP attach:

-- VectorCode Integration on LspAttach (if VectorCode is loaded)
if package.loaded["vectorcode"] and client.name ~= "vectorcode-server" and client.root_dir then
    local vc_ok, cacher = pcall(require, "vectorcode.cacher")
    if vc_ok then
        -- Check asynchronously if vectorcode is configured for this project root
        -- cacher.async_check("config", function()
        cacher.utils.async_check("config", function()
            -- Register buffer only if enabled and root_dir exists
            if client.root_dir then
                if cacher.lsp.buf_is_enabled(bufnr) then
                    cacher.lsp.register_buffer(
                        bufnr,
                        { project_root = client.root_dir, n_query = 10 } -- Adjust n_query as needed
                    )
                    -- vim.notify("VcodebaseectorCode registered for buffer: " .. bufnr .. " in " .. client.root_dir, vim.log.levels.DEBUG)
                end
                if cacher.default.buf_is_enabled(bufnr) then
                    if cacher.default.buf_is_enabled(bufnr) then
                        cacher.default.register_buffer(
                            bufnr,
                            { project_root = client.root_dir, n_query = 10 } -- Adjust n_query as needed
                        )
                        -- vim.notify("VcodebaseectorCode registered for buffer: " .. bufnr .. " in " .. client.root_dir, vim.log.levels.DEBUG)
                    end
                end
            end
        end, function()
            vim.notify("Failed VectorCode config check for " .. client.name, vim.log.levels.WARN)
        end)
    end
end

Neovim version: 0.11.4
OS: MacOS 15.6.1, Apple Silicon (M3)
I didn't notice anything wrong in the logs.
My nvim VectorCode config is the same as minusfive.

[Davidyz]

@Bakrog did you get the same error traceback? If not, please include it here to help me locate the problem.

[Bakrog]

Sorry, I forgot, here it is:

   Error  21:38:27 msg_show.lua_error

Error executing vim.schedule lua callback: ...Cellar/neovim/0.11.4/share/nvim/runtime/lua/vim/iter.lua:1067: src must be a table or function
stack traceback:
	[C]: in function 'error'
	...Cellar/neovim/0.11.4/share/nvim/runtime/lua/vim/iter.lua:1067: in function 'iter'
	...hare/nvim/lazy/VectorCode/lua/vectorcode/cacher/init.lua:23: in function ''
	vim/_editor.lua: in function <vim/_editor.lua:0>