Issue with extensions parameter ?

[smutel]

Hello,

I put this in my lychee configuration:

extensions = ["md"]

However when I start lychee, I have this answer:

[utils/test.py]:
     [404] https://docs.aws.amazon.com/aws-managed-policy/latest/reference/%7Barn.split('/')[1]%7D.html | Rejected status code (this depends on your "accept" configuration): Not Found

What I mess ?

[thomas-zahner]

Hi there, could you provide the relevant line/lines of the utils/test.py file with the link that is being checked?

[smutel]

            managed_policies = []
            for arn in policies.get("ManagedPolicies", []):
                managed_policies.append({
                    "Arn": arn,
                    "Description": self.get_managed_policy_description(arn),
                    "DocumentationLink": f"https://docs.aws.amazon.com/aws-managed-policy/latest/reference/{arn.split('/')[1]}.html"
                })

            data = {
                "PermissionSetName": name,
                "PermissionSetArn": perm_set_arn,
                "Accounts": [self.get_account_name(account) for account in accounts],
                "ManagedPolicies": managed_policies,
                "InlinePolicy": policies.get("InlinePolicy")
            }

[thomas-zahner]

Ah, I think I misunderstood your question initially and got sidetracked. Your question probably is; why is lychee checking the python file, if you include extensions = ["md"].

The answer is that the file arguments take precedence over configured file extensions. You probably run lychee */** or something equivalent. */** is an example glob which selects the utils/test.py file. The specified extensions only have an effect on directory traversal and directory traversal does not happen if files are specified directly/explicitly. lychee */** is expanded (either by the shell if unquoted or by lychee) to listing every single file.

In other words lychee --extensions md . does not check utils/test.py but lychee --extensions md ./utils/test.py does. So try something like lychee . or lychee src where there is no glob which expands to the python file.

Btw, you could also try to exclude the specific link which lychee doesn't properly detect or the single file. This way you could still do link checking in your python files.

[smutel]

Thank you for your answer @thomas-zahner. Make sens !
I am using lychee with pre-commit and a specific config file.
Here is my pre-commit config:

- repo: https://github.com/lycheeverse/lychee.git
    rev: lychee-v0.22.0
    hooks:
      - id: lychee
        args: ["-c", ".lychee.toml"]

What is the best to do what you suggest in your previous message ?
Remove the option extensions from the config file and put it in the pre-commit config like below:

- repo: https://github.com/lycheeverse/lychee.git
    rev: lychee-v0.22.0
    hooks:
      - id: lychee
        args: ["--extensions", "md", -c", ".lychee.toml"]

[mre]

Add this to your .lychee.toml:

exclude = [
  '^https://docs\\.aws\\.amazon\\.com/aws-managed-policy/latest/reference/.*'
]

Here is an example TOML file.

[smutel]

Here is the pre-commit configuration I am using:

  - repo: https://github.com/lycheeverse/lychee.git
    rev: lychee-v0.22.0
    hooks:
      - id: lychee
        args: ["--extensions", "md", "-c", ".lychee.toml", "."]
        pass_filenames: false

[mre]

I don't understand. Does your lychee configuration file exist? So, does .lychee.toml exist? If so, put this inside:

exclude = [
  '^https://docs\\.aws\\.amazon\\.com/aws-managed-policy/latest/reference/.*'
]

That's all you have to do to make it work. Sorry, but without any further information, we can't help you.

[smutel]

Oh sorry the previous pre-commit configuration is working fine to focus on markdown files without setting this exclude exception.

[mre]

Ah, got it.