Using the 22 Mar beta 1 build, this page http://www.fool.com/news/2000/foth000322.htm displays the main article in what looks like 24 point size. The same page displays the font size normally in 4.7

Harish, please track this down.

What I see ( refer to the attachment ) is the text "A Post Mortem on MicroStrategy" being rendered differently in Gecko. That is, the text appears bigger in Nav 4.x than in Gecko.

Content model ( genereated form the reduced case ): docshell=0110B7A0 html@027AA788 refcount=3< head@027AA728 refcount=2< Text@0278EC10 refcount=2<\n > style@0278DDE8 refcount=2< Text@0278DD20 refcount=2<\r\n .headline {font-family:arial,helvetica;font-size:150%;}\r\n > > > Text@02789D00 refcount=3<\n > body@02789C98 refcount=3< Text@02436680 refcount=3<\n > font@02439E18 color=red size=-1 refcount=4< Text@0247A360 refcount=5<\n Font size -1\n > font@02479768 color=blue size=-2 refcount=4< Text@024798D0 refcount=3<\n > br@02479878 refcount=3<> Text@0247C9E0 refcount=3<\n Font size -2\n > > Text@0244F660 refcount=3<\n > > p@02461D38 refcount=3< Text@02462070 refcount=3<\n > span@024640B8 class=headline refcount=3< Text@02467FB0 refcount=3<\n > font@02467548 color=red size=-1 refcount=3< b@0246E728 refcount=3< Text@02469F00 refcount=3<\n A Post Mortem on MicroStrategy\n > > Text@02469E90 refcount=3<\n \n > > > > Text@0245F240 refcount=3<\n> > > Looks like the parser is doing the right thing. The problem, however, seems to be style sheet related. CC in pierre.

Looks like this bug is a spin-off from bug 25810. Reassigning to attinasi@netscape.com

Harish forgot to say that my testcase displays big characters in Nav4x and Moz from 1 week ago, and small characters in IE and Moz from more recent builds. It certainly corresponds to the quirk you removed for bug 25810. However that quirk isn't particularly important. The problem we should solve is the one that appears on fool.com (see URL above) and in the "Reduced case" from Harish.

I don't think this related to removing the quirk in bug 25810 since it happens in standard mode... (25810 was a quirk-only change). Anyway, I'll look into why the <FONT size=-1> is causing the font size to be smaller than it seems it should.

Forgive me, Pierre, I spoke too soon! The problem is indeed the quirk, as you implored, and the removeal of the quirk is making the <FONT size=-1> get applied in Quirk mode and in Standard mode. As for the testcase: the <FONT size=-1> is getting applied correctly. Prior to my removal of the NavQuirk which would have caused the <FONT size=-1> tag to be ignored, the text would have been LARGE (350%). Now it is small (3-1=2 which is 2 pts. smaller than the default font size for the platform. Given that we axe'd the Quirk I say the testcase is OK. (The HTML4 spec. (section 15.2.2) indicates that the font size should be relative to the BASEFONT, so this is actually the correct behavior.) As for the fool.com site, the problem is a missing </FONT> tag. A reduced testcase of the problem will be attached showing that a missing </FONT> is causing the problems. Harish, can you help with the missing font tag? Note that the problem exists in Standard and Quirk mode (change the DTD on the new testcase to see). I'll also check in the actual fool page with a comment where the /FONT should be.

Corrected summary: the problem is that Gecko makes the font size smaller, no larger.

Just in case: last week I went across a testcase that you may want to play with because it shows a wide range of negative and positive font sizes. The testcase is at http://bugzilla.mozilla.org/showattachment.cgi?attach_id=6654. It was part of #32321 which, I think, isn't related to the current bug.

Moving to M17 - Harish, have you looked at the missing FONT tag problem?

I'll take it back...wonder why I pointed to style...let me figure that out.

After a bit of looking and discussing w/ Harish, it looks liek a font size problem after all. Taking it back.

The FOOL page looks fine right now. There are some problems with how residual style is kicking in and how the CSS rule for font size is conflicting, but we cannot really change how that works. Anyway, the headline is large as it is suppossed to be. I'm marking WORKSFORME since the page is rendering fine. For those interested, the Residual Style Problem: For the following HTML <FONT size=-1>F1 <FONT size=-2>F2</FONT> <P> <SPAN class=headline> HEADLINE </span> residual style will kick in and create this: <FONT size=-1>F1 <FONT size=-2>F2</FONT> </FONT> <P> <SPAN class=headline> <FONT size=-1>HEADLINE</FONT> </span> The insertion of the FONT tag inside of the span is causing the font-size specified in the style rule for class=headline to be overriden. Putting the FONT tag in front of the SPAN would fix it, but then other pages would break... Residual Style problems are hard, good thing this one is rendering fine as-is.

Verified Win98 2000092808