tidy converts '&' in query parameters (&aen=true => &aen=true) in relative paths

[JESii]

Input HTML

<a class="m-j-cm-link" href="/shop/womens-clothing?id=118&edge=hybrid&aen=true&cm_sp=mew_us_homepage-_-slot1-_-Women">

Output tidied HTML

"/shop/womens-clothing?id=118&edge=hybrid&aen=true&cm_sp=mew_us_homepage-_-slot1-_-Women">

I couldn't see an option to turn off this behavior...

[geoffmcl]

@JESii I do not think there is a tidy option to turn this off...

We do have an option --fix-uri no, but this only effects the escaping of illegal characters, and does not effect an ampersand...

But why would you want it turned off?

If you view the link as tidy has escaped it, in a browser, it should appear without the &, just with the &, so the link will function correctly...

Maybe I misunderstand the case... need further feedback... thanks...

[mc0e]

Tidy's behavior here is turning incorrect html into correct html. I'd have to say though that I've never liked the readability of those & character entities, and I much prefer the use of ;, which is blessed by the CGI spec as an alternative to the ampersand in query strings.

[JESii]

Well... I've never seen it converted to & in the href attribute; in fact the wikipedia entry on query strings indicates that either '&' or ';' (as mentioned by @mc0e above) is allowed. I'm not worried about viewing the ampersand; I want it to just be the delimiter for and the next key/value pair. it may view better, but from a programming perspective, it makes it harder to see what's going on -- the ';' is easy to miss (I've made that mistake before :=)

[mc0e]

& is correct in the query string, but to represent an & in html, you write it as a character entity. With html, an unescaped & is usually not the start of something that looks like a defined character entity, and browsers deal with that. In XHTML though it violates the requirement of being well formed XML, and the browser is required to reject it.

[JESii]

OK... got it. thanks!