From: David Abrahams (dave_at_[hidden])
Date: 2002-10-08 19:43:30
Beman Dawes <bdawes_at_[hidden]> writes:
> At 05:15 PM 10/8/2002, David Abrahams wrote:
> >> Sure it makes sense, but the net result was that a link which worked
> >> with the two most common browsers (IE and Netscape) no longer worked.
> >I don't see what point you're trying to make. I'm using Tidy as a
> >validity checker/fixer. Configured the way it is, it's doing its job
> >perfectly. There's nothing scary to me about the fact that it says
> >"this link is invalid, I'm changing it to something legal". I could
> >turn that check off, but then I would be writing invalid HTML.
> Turning invalid, but working, HTML into valid, but non-working, HTML
> doesn't seem like a practical approach.
Never said it was (note that Tidy has no way of knowing whether the
link has a target in the first place, or whether the valid or invalid
URI is the "real" one).
> It seems to me the solution is to apply the fix to both source and
> target files. Applying to only one or the other just results in links
> broken for practical purposes, even though the links are theoretically
Which is why I didn't check in a change with that fix in it.
> Applying the fix to both files results in links with are both valid,
> in the sense of meeting specs, and working, in the sense that browsers
> can use them as intended. One without the other isn't helpful.
Yes, but if other pages are linked to that page, I've broken
them. That's why I was trying to grep for all instances of spaces in
URIs. When I failed at that, I gave up.
-- David Abrahams * Boost Consulting dave_at_[hidden] * http://www.boost-consulting.com
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk