This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
When resolving relative links such as <img src="img/vccoursepg.gif" /> the link checker appears to use the document being checked as the base from which to reconstruct the complete URL, instead of the base specified in the <head> of the document using the <base href="" /> element (in those cases when the latter is present). This causes the link checker to create non-existent URLs, about which it then proceeds to complain. Example URL of document with this problem: http://academic.venturacollege.edu/mbowen/courses/2011haru/grading_status.shtml Base specification coded within this page: <head><base href="http://academic.venturacollege.edu/mbowen/index.shtml" /></head> Example of a link from this page that works in browsers but is flagged by Linkchecker (I will call this the "tail"): <img src="img/vccoursepg.gif" /> URL which Linkchecker *should* be constructing to test (base directory + tail), and which my browser (Firefox 3.x, 4.x) also successfully constructs: http://academic.venturacollege.edu/mbowen/img/vccoursepg.gif Nonexistent URL which Linkchecker actually constructs to test (directory of source page + tail): http://academic.venturacollege.edu/mbowen/courses/2011haru/img/vccoursepg.gif
You are right, this appears to be a regression introduced in link checker version 4.7 and still present in 4.8 - it works in version 4.6 in my local tests. Thank you for the report, will look into it ASAP.
Should be fixed now in the link checker development instance running at http://qa-dev.w3.org/wlc/checklink