relative broken links not detected

[lenisha]

Relative Links in the github pages index.html are broken but not detected:

https://github.com/lenisha/linkchecker-test/runs/2122896789?check_suite_focus=true

+ lychee '--exclude=^(javascript|chrome):.*' --verbose --accept=200,403,429 --base-url https://lenisha.github.io/linkchecker-test/   https://lenisha.github.io/linkchecker-test/ --output /tmp/lychee/out.md
+ exit_code=0
+ '[' 0 -ne 0 ']'
+ cat /tmp/lychee/out.md
+ echo
+ echo ::set-output name=exit_code::0
📝 Summary
---------------------
🔍 Total...........17
✅ Successful......17
⏳ Timeouts.........0
🔀 Redirected.......0
👻 Excluded.........0
🚫 Errors...........0

while index.thml has broken link https://github.com/lenisha/linkchecker-test/blob/main/index.html#L53

<a class="nav-link pr-3" href="./abouterror.html">About</a>

that points to https://lenisha.github.io/linkchecker-test/abouterror.html

Could it be also checked locally? Similar to discussion #15

[lenisha]

@mre is there any option to detect problems with relative links right now?
I tried to go against the live website with base-url

lychee '--exclude=^(javascript|chrome):.*' --verbose --base-url https://lenisha.github.io/linkchecker-test/   https://lenisha.github.io/linkchecker-test/ 

and tried against files files in PR

lychee  --exclude="^(javascript|chrome):.*"   --accept=200,403,429 "**/*.md" "**/*.html"

non have found the simple error

<a class="nav-link pr-3" href="./abouterror.html">About</a>

[pawroman]

@mre is there any option to detect problems with relative links right now?

There should be, once #99 is merged. Right now, link-parsing debugging is a bit cumbersome (e.g. you have to modify the source).

[mre]

No clue why it doesn't find the relative links, though. @pawroman do you know why?

[lenisha]

@mre @pawroman weird thing is when I explicitly point to index.html page it finds it

lychee --verbose '--exclude=^(javascript|chrome):.*' --accept=200,403,429 --base-url https://lenisha.github.io/linkchecker-test/ https://lenisha.github.io/linkchecker-test/index.html --output /tmp/lychee/out.md
+ exit_code=2
+ '[' 2 -ne 0 ']'
++ dirname lychee/out.md
+ mkdir -p lychee
+ echo 'Errors were reported while checking the availability of links.'
+ echo
+ echo '```'
+ echo
+ cat /tmp/lychee/out.md
+ echo
📝 Summary
+ echo '```'
+ echo
+ echo '[Full Github Actions output](https://github.com/lenisha/linkchecker-test/actions/runs/662263907?check_suite_focus=true)'
+ cat /tmp/lychee/out.md
+ echo
+ echo ::set-output name=exit_code::2
---------------------
🔍 Total...........28
✅ Successful......27
⏳ Timeouts.........0
🔀 Redirected.......0
👻 Excluded.........0
🚫 Errors...........1

Errors in https://lenisha.github.io/linkchecker-test/index.html
✗ https://lenisha.github.io/linkchecker-test/abouterror.html [404 Not Found]

also wonder why --verbose does not work in this combination

and local files could not detect it even with explicit file pointer:

+ lychee '--exclude=^(javascript|chrome):.*' --accept=200,403,429 index.html 
+ exit_code=0
📝 Summary
+ '[' 0 -ne 0 ']'
+ cat /tmp/lychee/out.md
+ echo
+ echo ::set-output name=exit_code::0
---------------------
🔍 Total...........17
✅ Successful......17
⏳ Timeouts.........0
🔀 Redirected.......0
👻 Excluded.........0
🚫 Errors...........0

[mre]

After thinking about this some more I think I know what's up:
The html file extension triggers the HTML mode in lychee. URLs without any file extensions get treated as plaintext, in which case relative URLs don't exist (currently at least).
We can probably safely assume that a URL returns HTML by default, so we have to change the condition here to return FileType::HTML by default if there is no extension:

lychee/src/collector.rs

Line 141 in ae2d02b

file_type: FileType::from(url.as_str()),

[mre]

@lenisha, this is fixed in master now. Will publish a new version of the lychee-action (1.0.7) shortly.