Created attachment 529230 [details] test case If I set display:inline on a <legend>, that should pull it back out of the associated <fieldset>'s box border. (Ideally, there would be a special display: value that has the effect of "put this into the containing element's box border" and setting display: to *anything else* should cancel that. I'm aware that I *can* pull a <legend> out of the box border with float:, but that still doesn't make its text baseline-aligned with other text in the same form.) Rationale: I might want to use <legend> and <fieldset> strictly for their semantic consequences, such as their effect on screen readers, see http://www.paciellogroup.com/blog/2007/11/fieldsets-legends-and-screen-readers/ -- and not their layout effects. In the test case I've attached, this would be one of several multiple-choice questions in a form, so inline layout would be much more appropriate -- note also that the <fieldset> border has been disabled. IE 8 and 9 do what I (my informant) wants. Webkit is the same as Firefox.
Further exposition on why this shouldn't be either a dupe or a dependency of bug 292736: that bug is substantially more ambitious than this one. It would be *nice* to be able to duplicate the effect of <legend> inside <fieldset> with nothing but CSS, but being able to *turn the magic off* is all I need for this bug, and I'm betting that's a lot easier to implement. Let's not wait a straightforward, desirable, well-scoped improvement on a much more nebulous and ambitious one.
I'm not sure how web-compatible this is. I seem to recall I kept the current behavior when we were making <legend> more stylable because there were pages out there that were using "display: inline" on legends and expecting that to keep them as legends.... Which IE mode(s) was the testing in comment 0 done in? Does the behavior depend on mode?
Amusingly enough, btw, I did just see http://stackoverflow.com/questions/5818960/legend-not-displaying-inline go by.
(In reply to comment #4) > I'm not sure how web-compatible this is. I seem to recall I kept the current > behavior when we were making <legend> more stylable because there were pages > out there that were using "display: inline" on legends and expecting that to > keep them as legends.... It would be okay by me if there was something else you had to do as well as setting display:inline, as long as it was a documented (and hopefully at least submitted for standardization) something. > Which IE mode(s) was the testing in comment 0 done in? Does the behavior > depend on mode? With IE9, in its default mode and its IE8 compatibility mode, display:inline pulls the legend out of the border. IE9 in IE7 compatibility mode, however, does not. My informant didn't try any other modes or any actual installs of IE8 or 7, and I'm on a Mac right now :)
Is IE7 compat mode the "quirks" mode in IE9? Or is it one of the standards modes?
I think quirks mode in IE9 is orthogonal to the IE7 and 8 backward compatibility modes.
You can get immediately IE5.5 - IE9 screenshots oft testcases here: http://meineipadresse.de/netrenderer/index.php (replace https:// with http://) > Is IE7 compat mode the "quirks" mode in IE9? it's one of the standards modes
And "inline" seems to be legend's default in IE quirksmode. "block" expands width to 100%.
This still seems to be an issue in FF. Ran into this issue here: https://drupal.org/node/2236789 I'm a bit surprised that this isn't an issue that has occurred more often.
(In reply to Mike Gifford from comment #11) > This still seems to be an issue in FF. Ran into this issue here: > https://drupal.org/node/2236789 > > I'm a bit surprised that this isn't an issue that has occurred more often. http://jsfiddle.net/yc70ptdq/1/ Firefox now seems to display it inline. If you add a surrounding <fieldset> element, however, it snaps back to the old display.
(In reply to Boris Zbarsky [:bz] from comment #4) > I'm not sure how web-compatible this is. Note that I think Blink has made this change for version 50 in https://bugs.chromium.org/p/chromium/issues/detail?id=481983 so we can see if they revert due to compatibility issues.
That's hilarious. They made the change ostensibly to match us, but implemented a totally different behavior from ours.... And in the process started clearly violating the HTML spec.
Oh, and given that we're following the spec here, marking invalid. If people want the spec to change, the right approach is open discussion on the spec, not guerilla behavior changes by browsers...
Status: NEW → RESOLVED
Last Resolved: 3 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.