How to --dump to file?

[tooomm]

It looks like --dump only prints the found links to the console.

I would like to dump the extracted links to a file to further work with it.
Not sure how to do this with the action.

output: is not considered, or I'm not sure of the actual location 🤔

[mre]

Yeah, that's a bug in lychee then. It should consider the output setting. Let me move that over to the main repo.

[mre]

Actually I'm confused. It should work:

https://github.com/lycheeverse/lychee/blob/11d8d448953d85300f2a5b22815cc0d310bd7ff2/lychee-bin/src/commands/dump.rs#L21C9-L30

This has always been the case, ever since we supported dumping links in the first place.
lycheeverse/lychee@35ccfb8

[tooomm]

Hmm, did you make it work with you GH Action?
As I noted, I might just don't know where the file ends up and how to properly call it from the action.

If you've an example at hand that would be awesome!

[mre]

Whoops, then let me move it back to lychee-action. Apologies for the confusion.
You can decide where the file ends up in the action.

- name: Link Checker
  uses: lycheeverse/lychee-action@master
  with:
    # Check all markdown and html files in repo (default)
    args: --base . --dump './**/*.md' './**/*.html' './**/*.rst'
    # Use different output file path
    output: /tmp/foo.txt

The default path currently is lychee/out.md (as defined in the action.yml).
If the output does not end up there, that's a bug.

To reuse the output in a downstream job that depends on this job, see "Defining outputs for jobs" in the GitHub Docs.

[tooomm]

Thanks again for trying to help here, mre!

You can decide where the file ends up in the action.

That's correct.
As stated in the issue description, I was using output and was not able to spot the file when trying various possibilities.
Same actually, when using the default location and not using a custom path for output.

So the question is still the same... :D
Can not tell if it's a bug or me being the issue here.

To reuse the output in a downstream job that depends on this job, see "Defining outputs for jobs" in the GitHub Docs.

It should only be used in a subsequent step within the same job.

Can you provide me a working example that e.g. prints the file afterwards in a following step in the action?
Taking your above example here, how to print the foo.txt file for example... I could not find/link the file in my tests.

jobs:
  test-job:
    runs-on: ubuntu-latest
    steps:
      - name: Dump Output
        uses: lycheeverse/lychee-action@v1
        with:
          # Check all markdown and html files in repo (default)
          args: --base . --dump './**/*.md' './**/*.html' './**/*.rst'
          # Use different output file path
          output: /tmp/foo.txt
    
      - name: Print Output
        shell: bash
        run: cat ???

[mre]

Hum, I don't know the answer. This doesn't work at least:

name: Links

on:
  workflow_dispatch:

jobs:
  linkChecker:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Dump Output
        uses: lycheeverse/lychee-action@v1.8.0
        with:
          args: --dump './**/*.md' './**/*.html' './**/*.rst'
          output: ./lychee/out.md
          debug: true

      - name: Print Output
        run: cat ./lychee/out.md

Output:

Run cat ./lychee/out.md
cat: ./lychee/out.md: No such file or directory
Error: Process completed with exit code 1.

I've added the debug flag. Maybe that helps to troubleshoot the issue.

[joryirving]

So I'm running into a similar situation, where when I pass in the output command, it fails: https://github.com/LilDrunkenSmurf/k3s-home-cluster/actions/runs/7309944973/job/19918081978#step:4:110
I setup debug, and tried to print the output, and I noticed it's not evaluating the second if statement here:

lychee-action/entrypoint.sh

Line 33 in ef8c8f3

if [ -f "${LYCHEE_TMP}" ]; then

It looks like it runs the if statement above, , and evaluates it to false, and carries on, but skips the next one.

Rather than making it two different if statements, perhaps it should be nested.

Something like:

# Ensure we have any output
if [ ! -f "${LYCHEE_TMP}" ]; then
    echo "No output. Check pipeline run to see if lychee panicked." > "${LYCHEE_TMP}"
else if
	# If we have any output, create a report in the designated directory
    mkdir -p "$(dirname -- "${INPUT_OUTPUT}")"
    cat "${LYCHEE_TMP}" > "${INPUT_OUTPUT}"

    if [ "${INPUT_FORMAT}" == "markdown" ]; then
        echo "[Full Github Actions output](${GITHUB_WORKFLOW_URL})" >> "${INPUT_OUTPUT}"
	fi
fi

[mre]

Cool. Can you send a pull request with that change, so we can test it?

[joryirving]

Tested the PR, and it looks good.

[mre]

@LilDrunkenSmurf fixed it in #215. Thanks!
We've also added an integration test in #217 to make sure it doesn't break in the future.
@tooomm fyi.

[mre]

Also added an integration test for --dump specifically. #218