Webmaster wrote:
Georg, the fix doesn't suggest putting different values on html and
body (or did I miss the whole point?).
I understood the solution to be setting body and/or html to 100.01%
and then setting any other styles and text-level attributes with ems
or %. Did I get it wrong?
Not necessarily, but the fix for that particular IE-bug is to define
font-size in percentage on the highest level, html *or* body *or*
'another element', and not to define font-size in em above it. This can
be achieved in (at least) two ways.
Therefore:
html {font-size: 100%;}
body {font-size: (whatever-value)em;}
...is a perfect bug-fix with cross-browser reliable results.
and the usual:
html {/* no font-size defined at all */}
body {font-size: (whatever-value)%;}
...is also a perfect bug-fix with cross-browser reliable results.
My point was/is that any other value-combination that can be calculated
to end up as the "right" base-size will also fix the bug -- but seems to
only give the calculated results in IE/win.
---
My reason for mentioning the whole thing is that there are always the
possibility that some may try to be "smart" and rearrange those values
in the first example without running a thorough cross-browser test.
They may end up with a perfectly valid "human bug" in their page-code--
banging its head against against some nasty "human bugs" that ended up
as browser-code, and such "wars of logic" can give pretty unpredictable
endings. So, reason enough to mention it.
Sorry if I complicated the matter for some, but after 25 years of
trouble-shooting solutions around buggy software, the tendency to go in
depth is irresistible. Must be a "human bug"-- somewhere :-)
---
I've just done a complete run of screenshots using the above method
and can still see some variations in size, particularly at larger
sizes (h1-h3).
Minor variations are usually caused by browsers having different
'tip-over' values for recalculating em/% into physical screen-units.
Some 'tip over' between .99 and .00, while other 'tip over' around
.40-.50. A few levels of inheritance and such differences can become
quite noticeable, but they are more or less unavoidable--apart from by
keeping the number of levels low.
Aside from such minor variations; I got some really huge variations
during my own testing with different 'html+body' values. My need for
killing "human bugs" during any debugging-process is why I asked for
confirmations or corrections.
---
For the record: my own "extreme font-resizing bug" fix is to feed
standard-compliant browsers a 'pixel-defined font-size on body', and
feed IE/win a 'percentage-defined font-size on body' through a
'conditional commented' stylesheet that I'm using anyway. That seems to
work quite well across browser-land -- and the bug is dead.
Georg
--
http://www.gunlaug.no
******************************************************
The discussion list for http://webstandardsgroup.org/
See http://webstandardsgroup.org/mail/guidelines.cfm
for some hints on posting to the list & getting help
******************************************************
