Hilko Bengen wrote: > I think that throwing away aggregated content if tidy couldn't cope with > it is not right either. Such behavior annoys users a great deal and it > is likely to confuse them because the "bug" seems unreproducible due to > varying external inputs over which the user has no control.
I actually don't think that's as annoying as having one bad aggregated
item break display of everything after it. There are lots of other ways
that aggregated content can fail to display in ikiwiki -- for example,
it might rely on embedding a youtube video, which htmlscrubber would
disallow.
Tidy has fairly few things that trigger actual errors; one of them is this:
<a href="foo>bar</a>
1
2
3
4
5
6
7
8
9
10
11
The missing quote will cause some browsers (those without tag soup)
to ignore the entire rest of the page up to the next quote. And
htmlscrubber, which tends to be used along with htmltidy, has to parse
the html like that, so will scrub it into something unrecognisable even
if tidy leaves it alone.
The only solution I can think of that is better than throwing such
unparseable html away is stripping all html markup from it. But if you
try to strip all html markup, it's hard to find a way to do it that
doesn't turn the above example into "".
--
see shy jo
signature.asc
Description: Digital signature
