Eric Van Dewoestine wrote:
Is there a particular reason why <xxx/> tags are skipped when using
"at" (:h at)?

The tag-blocks documentation notes that this is by design but offers
no reasoning behind it.  For html where the only self closing tags are
usually <p/> or <br/>, I could see how this behavior might be
justified.  However given a larger xml tag:

<sometag attr1="one" attr2="two" attr3="three" attr4="four"/>

It would be very helpful if "at" could be used for selection.

Thoughts?


at and it skip not only <xxx/> self-closing tags, they also skip <p> and <li> when they don't have a corresponding </p> or </li> (<p/> is invalid in HTML), <meta ... > and <br> even when they aren't ended by />, etc.

The idea is that at "a tag block" needs means to define where a block begins and ends, i.e., a begin tag and an end tag. A self-closing tag has by definition no "block of text", no "content" that it could select.

With the cursor within a self-closing tag such as, let's say, <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />, which is typical for HTML and not "trivially short", you could select the tag itself with a<

Outside the span of any tag, you can find the next tag with /<

The only problem is having at select e.g. from within an HTML paragraph, leftwards to and including the <p> tag, and rightwards to where a </p> tag ought to be but is missing, e.g. just before the <p> tag for the next paragraph, or just before a closing tag for an enclosing block-level tag such as </div>. Such "missing closing tags" are allowed in HTML but not in XHTML; but having at and it work like that would require a detailed knowledge of HTML syntax, and I think it is too complex a task for a "simple" command like it or at .

IMHO it is "the trend of the future", and therefore "good form" to include </p> at the end of paragraphs, </li> at the end of bulleted or numbered list items, </td> at the end of table cells, etc., and /> at the end of unpaired tags. Doing so also solves the problem mentioned in the above paragraph.


Best regards,
Tony.

Reply via email to