35 lines
1.3 KiB
Plaintext
35 lines
1.3 KiB
Plaintext
|
|
Extracting inline CSS from HTML Purifier
|
|
voodoofied: Assigning semantics to elements
|
|
|
|
Sander Tekelenburg brought to my attention the poor programming style of
|
|
inline CSS in HTML documents. In an ideal world, we wouldn't be using inline
|
|
CSS at all: everything would be assigned using semantic class attributes
|
|
from an external stylesheet.
|
|
|
|
With ExtractStyleBlocks and CSSTidy, this is now possible (when allowed, users
|
|
can specify a style element which gets extracted from the user-submitted HTML, which
|
|
the application can place in the head of the HTML document). But there still
|
|
is the issue of inline CSS that refuses to go away.
|
|
|
|
The basic idea behind this feature is assign every element a unique identifier,
|
|
and then move all of the CSS data to a style-sheet. This HTML:
|
|
|
|
<div style="text-align:center">Big <span style="color:red;">things</span>!</div>
|
|
|
|
into
|
|
|
|
<div id="hp-12345">Big <span id="hp-12346">things</span>!</div>
|
|
|
|
and a stylesheet that is:
|
|
|
|
#hp-12345 {text-align:center;}
|
|
#hp-12346 {color:red;}
|
|
|
|
Beyond that, HTML Purifier can magically merge common CSS values together,
|
|
and a whole manner of other heuristic things. HTML Purifier should also
|
|
make it easy for an admin to re-style the HTML semantically. Speed is not
|
|
an issue. Also, better WYSIWYG editors are needed.
|
|
|
|
vim: et sw=4 sts=4
|