Hacker News Reader

Comment by kleiba | Hacker News Reader

OT: in Emacs, you could prevent this from happening by using C-u M-% (replace-word): in that case, only those instances of "div" that are separate words (not parts of other words) will be replaced.

How would you avoid "inpiduals" (and similar mistakes) in other editors?

View on HN

"Match whole words" is pretty common. And in every editor invariably ends up disagreeing with you on what precisely constitutes a word boundary.

In this particular case I would just have included the starting < in the search and replace.

You mean div> otherwise it wouldn't pick up the closing tags :)

I believe replacing `<div`\'<p' and `/div`\'/p' are preferred so you can retain any tag properties (class, id, style, etc), but it could still leave behind any tags with superfluous spaces.

If you replace div> then you will miss any opening tags that have attributes. E.g. <div class=baz> or <div id=foobar>.

In Vim I would do

    :%s@\(</\?\)div@\1p@g

It’s a little bit convoluted but it replaces the tag names in both opening and closing tags and does not replace it anywhere else.

Fair point...

In many editors, you can use a regex match for word boundaries on either side of div, like \bdiv\d. Some also have a checkbox for the word matching functionality.

In vim: :%s/\<div\>/p/g (add C at the end if you want to confirm each change)

Huh - that’s confusing in the context of this example, but indeed tests fine on my system (nvi).

Since the “word” in question is “<div>”, I thought the \< was just an escape. I looked in various docs and the best I see in Tcl re_syntax(n) is \< and \> are synonyms for [[:<:]] and [[:>:]] respectively, which are constraints matching empty strings at the beginning and end of words, where a word a sequence of word characters... [a-zA-Z0-9_]

TIL, too...

Right? Totally nonintuitive! But easy enough to remember and type. I figured I wasn't the only one who found it a bit surprising ^_^

TIL what \< means. Thank you.

Open it in a web browser and then copy and past it into the editor as plain text.

the underlying regex libraries often (like your emacs) have a symbol for word boundaries. emacs is great, though, agreed.

Just replace div> with p> ?

Reply on news.ycombinator.com