Title Links Not Working in 1.0.14
Closed this issue ยท 7 comments
In the latest release, I notice links to title sections are now marked as broken.
ie.
[Dependencies](#Dependencies)
# Dependencies
This markdown example has no dependencies.
I can work around this with configuration, but I wanted to check if this was the desired behaviour?
Thank you @niall-byrne, Can you verify this with v1.0.15 ?
@gaurav-nelson: I can confirm this does indeed happen with v1.0.15
. After seeing my prior issue #170 was fixed in v1.0.15
, I went to point at that version in my production config only to find the linkchecks of anchor links like these were broken, so that's when I searched & found this issue. I've just reproduced it in my test repo https://github.com/philrz/linkcheck: The check for [Dependencies](#Dependencies)
passed ok when I had it pointing @v1
but it fails when pointing @1.0.15
. I notice that @niall-byrne's paste shows what looks to be incorrect syntax [Dependencies][#Dependencies]
(two sets of square brackets) but I have it failing with [Dependencies](#Dependencies)
.
@philrz I think the correct markdown anchor should be:
[Dependencies][#dependencies]
...
...
# Dependencies
This markdown example has no dependencies.
Because href
is always generated like that (d
vs D
). Its working fine on GitHub because they maybe using some scripts.
Can you recheck with this change?
@gaurav-nelson: Ah yes, thanks for spotting the importance of the capitalization. I did the proposed recheck and indeed now that it's a lowercase d
the checker @1.0.15
is passing with that test README. So, that's great!
However, I've still got a problem with my production markdown, and unfortunately it doesn't seem to be a problem with capitalization. I've copied the markdown into my test repo at https://github.com/philrz/linkcheck/blob/main/zng.md if you're up for taking a look. The output of the Actions run with verbose mode enabled shows:
FILE: ./zng.md
[โ] https://avro.apache.org/ โ Status: 200
[โ] https://en.wikipedia.org/wiki/Apache_Parquet โ Status: 200
[โ] https://developers.google.com/protocol-buffers โ Status: 200
[โ] https://kafka.apache.org/ โ Status: 200
[โ] #21-types-frame โ Status: 200
[โ] #22-values-frame โ Status: 404
[โ] #23-control-frame โ Status: 404
7 links checked.
ERROR: 2 dead links found!
[โ] #22-values-frame โ Status: 404
[โ] #23-control-frame โ Status: 404```
What's strange is that if you look in the doc you'll see these links all appear right next to each other:
* [types frame](#21-types-frame),
* [values frame](#22-values-frame), or
* [control frame](#23-control-frame).
And the section headers for each are:
### 2.1 Types Frame
### 2.2 Values Frame
### 2.3 Control Frame
And for some reason it thought the Types Frame one was ok but the other two were a 404, but I can't for the life of me see why they're treated differently. I'd appreciate any insight you might have. Thanks!
Thank you @philrz I ran through that file locally with markdown-link-check
and got the same result. I'm not sure what is wrong here, but this issue is with the underlying markdown-link-check
library and not the GitHub action. I suggest to close this one and create an issue at https://github.com/tcort/markdown-link-check/issues
Thanks @gaurav-nelson. I've opened tcort/markdown-link-check#250. Will see what they say.
I'll close this issue given the upstream library is being investigated.