During testing, I got a normative reference in an informative section which rendered like this

But what I was expecting?:
The word BARRRR to be in single line.
Disabling display: inline-block for .respec-offending-element fixes the issue. Or changing to display:inline also works.
Can you share the HTML that caused this?
Just add <p>A normative reference in informative section [[!payment-request]] and is illegal.
But we keep it as normative and emit a warning.</p>
in basic.html after Line 54 (https://github.com/w3c/respec/blob/develop/examples/basic.html#L54).
Basically enforcing normative reference in any informative section will do.
So weird... I'd like to leave display: inline-block if possible... but we should figure out why it's doing the crazy rendering. I've never seen that. Can you check in multiple browsers?
Sorry, my bad, I should have mentioned this occurs in Firefox. Just checked in Chrome - worked fine.
Ok, reproduced... the bug is that base.css sets a[href] to margin: 0 -1px 0; for some reason. What might fix it is to set margin: 0; on respec-offending-element.
sorry, just on a[href].respec-offending-element.
I should file a web compatibility bug on Firefox... Chrome and Safari render this properly.
For the note, Edge also renders properly.
Good finding @Swapnilr1!
Thanks @sidvishnoi.
This is really strange - I never thought it could be a bug in Firefox.
@marcoscaceres minimal example: https://codepen.io/sidvishnoi/pen/BENGNR
Thanks - added the reduced test case to the Firefox bug.
Should we patch this by setting a[href].respec-offending-element { margin: 0;} ?
This will fix it in Firefox, and Chrome is unaffected by setting margin to 0.
I've filed a bug on the W3C to remove the -1: https://github.com/w3c/tr-design/pull/176
@Swapnilr1 yes, your suggestion should work (it think). Please confirm.
Setting margin to 0 for .respec-offending-element a[href] fixes the issue in Firefox, and in Chrome causes no changes.
We can add this as CSS in some module, though I don't know which one would be most suitable. But now that you have opened https://github.com/w3c/tr-design/pull/176 do we need to patch it?
We can add this as CSS in some module, though I don't know which one would be most suitable.
Whichever file currently defines respec-offendng-element. I think that's respec2.css?
Just to be clear, we should temporarily patch it... and then remove our patch once w3c/tr-design#176 gets fixed.
Oh wow, it was a known issue from 5 months ago: https://bugzilla.mozilla.org/show_bug.cgi?id=1502610
Strange nobody filed a bug here when the issue was discovered in a ReSpec built specification.
I think the consensus was that it was a Firefox bug, so not worth fixing in ReSpec. However, given recent changes in Chrome, it seems Firefox was the one doing the right thing and Safari is wrong.
Most helpful comment
Filed https://bugzilla.mozilla.org/show_bug.cgi?id=1541102