See bug 254659 comment 9. Testcase is in the URL field. It looks like line layout always reflows "inline" frames with an unconstrainted vertical height, including images in a line. I _think_ this is a regression...
Really? I don't think we've ever had the ability to break inline frames vertically. I don't think that fits into our pagination model at all... if an inline image did return incomplete, we would wrap the pieces and give them independent x offsets, which is NOT what you want.
Hmm. Maybe I'm confused... let me test.
So this changed with bug 297537. I guess we broke horizontally before that, line-wrapped the second half if the image was wide enough (which in all the testcases it appears to have been), and then the image frame painted the right thing in, huh?
That's my guess, yeah.
Sad, but understandable.
Can we please re-open this bug as it's a long wanted feature and a flaw. Even IE does this.
Reopened at request of Tobbi on livechat. As stated in comment #6 IE8 now does this.
This is wontfix as filed (which is to create continuation frames, etc). If you want to file a generic bug about doing something sane with too-tall inline images when printing, fine. But don't hijack other bugs.