https://bugzilla.wikimedia.org/show_bug.cgi?id=54604
Krinkle <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #10 from Krinkle <[email protected]> --- (In reply to comment #9) > (In reply to comment #8) > > The CSS is pretty ridiculous. Let's not forget that CSS parses from the > > rightmost rule so something like a.external[href ^="news:";] can be a pretty > > expensive operation. > > [citation needed]. I don't see why would it be particularly more expensive > than > any of the hundreds of other rules we have. Assuming that the hypothetical > browser does indeed naively match from the right side always, [href^="news:";] > will immediately not match (most elements don't even have a href attribute) > and > the ten or so of such rules should not be noticeable. Afaik selector engines have quick lookup maps for IDs, classes and elements. In the case of ".mw-link-pdf" that is a very quick lookup. In the case of "#bodyContent a.external[href=..]" it is a lookup for elements of 'a', then for each the classname and the attribute is matched, and then it recursively does a parent check until it finds #bodyContent or the document element. In theory it could optimise if it had a quick lookup for elements having a certain attribute. But afaik they don't have that, and even if they did, I isn't more efficient per se, cause then the post-check would have to verify tag name and class name for each. Either way, none of this in particular is bad, it's the whole package of: * A very strong selector (#bodyContent) * A parent selector * Both element and class name (should never be used together) * Additional attribute matching and string operations on that * Hardcodes details that (other) stylesheets should not have to know or hardcode, thus causing maintenance issues (e.g. the fact that it is an <a>, and is inside #bodyContent. The container can change, and links might be wrapped for presentational reasons) Up sides of using a simple class: * Faster query (one could argue about how much, but it is certainly faster) * Support for older browsers (this is not a priority though, since all our grade A browsers support it, and it's a minor enhancement with graceful degradation) * Easier to extend from other components as there is no need to have any knowledge about implementation details (container element, tag name etc.) -- You are receiving this mail because: You are the assignee for the bug. You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikibugs-l
