DESCRIPTION: The CSS font-family property and the HTML <FONT FACE=""> do not work on Linux if there is no generic family (serif, sans-serif, etc.) at the end of the font list. This effects lots of major pages all over the web (FONT FACE). STEPS TO REPRODUCE: Load testcase (to be attached). ACTUAL RESULTS: First and third lines are Verdana and Arial, respectively. Second and fourth lines are a serif font. (I have Verdana and Arial installed on my RedHat 6.0 Linux.) EXPECTED RESULTS: The first two lines should be in Verdana, and the third and fourth should be in Arial. DOES NOT WORK CORRECTLY ON: * Linux, apprunner, 1999-08-16-08-M9 WORKS CORRECTLY ON: * Windows, apprunner, 1999-08-16-09-M9 * Linux, NN 4.6 ADDITIONAL INFORMATION: Fixing this would make lots of major websites look a lot better. :-) Thus marking as major.

Looks like a platform specifc font selection bug.

dbaron@fasharvard.edu: Can you do a xlsfonts |egrep "Verdana|Arial" on your system and attach the result into this bug report ? I suspect there are a uppercase/lowercase issue here...

Erik: Do you think this is a beta stopper If not, please assign to M14 or M15. Thanks

The command you give shows nothing, but it shows a lot if I change it to lowercase. However, I doubt this is the problem, since (as you can see from the screenshot) I really do get Verdana and Arial (the fonts are different, so it's not a generic sans-serif) when the "sans-serif" is at the end. *I* think this is DEFINITELY a beta-stopper. If you'd like me to give you a list of sites where this is a problem, I can. I suspect (rough guess) it shows up on half of the top 100. For example (just guessing a few sites): http://www.cnet.com/ http://www.aol.com/ etc. (although a few pieces of that could be residual style bugs, I suspect they show much better on Windows)

I downloaded http://www.aol.com/ and did the following search-replace: "Geneva, Arial" -> "Geneva, Arial, sans-serif" "geneva, arial" -> "geneva, arial, sans-serif" and it caused most of the fonts on the page to get better (and I think the rest weren't caught by those search-replace operations). I will attach a screenshot of the difference between 4.x and Mozilla on http://www.aol.com/.

Note that the screenshot shows some font-sizing bugs in Nav 4.6 - it has problems with font sizes when interacting with xfsft. However, the font faces show serious problems in Mozilla.

Will try to fix for M12.

what will happen if you put down p.one { font-family: Verdana, some-other-value-to-make-Verdana-not-the-last-one; } or p.one { font-family: verdana, sans-serif; }

My theory still holds. It's not the capitalization in the document, and it's not that the font can't be last. The generic fonts fully explain the behavior I see. Attaching additional test case and screen shot.

David, what happens if you use "font-family: Verdana, serif"? Does it matter _which_ generic font-family is on the end?

My memory is that it doesn't matter which generic is used. I'm 80% sure I tried serif in addition to sans-serif. However, I can't check right now.

Based on the screenshot in bug 15989: http://bugzilla.mozilla.org/showattachment.cgi?attach_id=2061 , it seems that some people see this bug and some don't, since it doesn't show up in the screenshot. Thus it may be particular to my configuration. I use the xfs+xfsft that comes with RH 6.0, but I've added some TrueType fonts into the font path. I also run my X-server at 91 dpi. If you want me to see if undoing either of these changes fixes the problem, I can look.

Changing the logical resolution back to 75dpi didn't change anything. Messing with the font stuff is harder. Let me know if you want me to give it a try.

This is now working for me, and does in builds going back to late August. It seems to have been fixed between 8-26 and 8-30. However, I'm wondering why I was still seeing it in early October.

dbaron, Can we close this bug now?

I guess. I can always reopen it if it's still happening.

Verifying fixed

Migrating from {css1} to css1 keyword. The {css1}, {css2}, {css3} and {css-moz} radars should now be considered deprecated in favour of keywords. I am *really* sorry about the spam...