12461 - larger doesn't work when font-size <= 10px

Status: VERIFIED FIXED

Description

[David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 (away December 22 - January 13) (if account gets disabled due to email bounces, ask a bugzilla admin to reenable it)]

DESCRIPTION:  The larger keyword for the CSS1 font-size property doesn't work
when the parent element's font size is less than or equal to 10px.  This
includes the keyword xx-small.

STEPS TO REPRODUCE:  Load attached test case.

ACTUAL RESULTS:  The font only gets bigger on the second and fifth lines.  The
1st, 3d, and 4th lines have the same font size throughout.

EXPECTED RESULTS:  On every line, each word should be bigger than the one before
it.

DOES NOT WORK CORRECTLY ON:
 * Linux, viewer, 1999-08-24-08-M10
 * Windows 95, apprunner, 1999-08-24-09-M10

WORKS CORRECTLY ON:
 * Windows IE 4.0

Comment 1

[David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 (away December 22 - January 13) (if account gets disabled due to email bounces, ask a bugzilla admin to reenable it)]

Created attachment 1386 [details]
testcase described in bug

Comment 2

[David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 (away December 22 - January 13) (if account gets disabled due to email bounces, ask a bugzilla admin to reenable it)]

The threshhold seems to vary slightly by platform, actually.  The 11px case (5th
line) doesn't work in Windows, but it does in Linux.  Probably 12px or 13px
would work... since x-small does.

Comment 3

[troy]

I think this must be a style system. Layout just uses the info we get from the
style system

Comment 4

[Peter Linss]

Pushing off non-beta 1 issues

Comment 5

[Pierre Saslawsky]

Reassigning peterl's bugs to myself.

Comment 6

[Pierre Saslawsky]

Accepting peterl's bugs that have a Target Milestone

Comment 7

[extremely]

The problem looks like it is for fonts of 9pt and below not 10px and below.
That would account for the system differences, different px=>pt conversions.

With default values (12pt Times) the cut off in the html table code comes out
at 9.24pt or 184.8 twips.

Look at
http://lxr.mozilla.org/seamonkey/source/layout/html/style/src/nsStyleUtil.cpp#1
57

you seek slowly downward till you find a returned font size that is smaller
than the current one then go back up one step. Bing! always the same size for
fonts smaller than 9.24pt.  Change the greaterthan to a equals test and you
fix this bug, I think.

Hope it helps.

Comment 8

[extremely]

*sigh* it is line 157

Also, you should use >= rather than == just in case the math gets "jumpy" and
skips a value.  There is a pow() in there and index is it's exponent. As index
gets higher/lower it will REALLY swing that value. for outlandish text sizes it
might get buggy with ==.

Comment 9

[extremely]

Verified Fix under linux against m11 codebase

Comment 10

[Pierre Saslawsky]

Pushing my M15 bugs to M16

Comment 11

[Hixie (not reading bugmail)]

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...

Comment 12

[ekrock's old account (dead)]

Bulk moving [testcase] code to new testcase keyword. Sorry for the spam!

Comment 13

[Hixie (not reading bugmail)]

Pierre, you might wish to have a look at this while you do all the other font
stuff, especially since there is a fix described in the bug.

Comment 14

[rickg]

Moving crufty m14-m15 bugs out to m16 for proper triage.

Comment 15

[Pierre Saslawsky]

The algorithm has changed a while back. Larger/smaller acts on the HTML sizes 
which are constrained to [0 - 7] but within that range, they do work for any 
size. This behavior is identical to IE's.

Marked M18. A solution would be to fall back to the old algorithm when the font 
size is out of range, even if we are in strict mode.

Comment 16

[Pierre Saslawsky]

This bug has been marked "future" because the original netscape engineer working 
on this is over-burdened. If you feel this is an error, that you or another known 
resource will be working on this bug, or if it blocks your work in some way -- 
please attach your concern to the bug for reconsideration. 

Comment 17

[timeless]

Um, so what needs to be done here?
I'm going to attach a patch that does a bit of code cleanup.

Justification for loss of future: css1
relnote2: if we don't fully support css we need to say so.
nsbeta3: css1

Comment 18

[timeless]

Created attachment 11455 [details] [diff] [review]
patch reduce possible rounding error; r=[e-mail]

Comment 19

[Brendan Eich [:brendan]]

Marc, Pierre: could one of you also review Josh's patch and (if it passes muster 
with you guys) help get it checked in, a=[e-mail]?  Thanks.

/be

Comment 20

[Marc Attinasi]

This bug is not nsbeta2+ and should not be checked in. We are trying to minimize 
changes and increase stability, not fix every small bug.

Also, reassigning back to Pierre: why was it remvoed from his queue and sent to 
'nobody' in the first place?

Comment 21

[David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 (away December 22 - January 13) (if account gets disabled due to email bounces, ask a bugzilla admin to reenable it)]

Netscape employees should not be fixing bugs that are not nsbeta2+ .  That
restriction does not apply to others.  See the top of tinderbox.

Comment 22

[Brendan Eich [:brendan]]

[e-mail] has stated this too, most recently in mozilla.seamonkey:
you're a mozilla.org module owner (you happen to work for netscape.com).  Part
of your job as module owner is to take good patches when they're ready to go
into the trunk (even if they're "too risky" for your employers' product).  Yes,
module owners may have to serve two masters; life's rough.  [e-mail]
will have to take steps if an owner freezes out good patches for too long (where
good patches are defined by the owner and peers in the entire community, based
on code review and technical merits).

In this case, I don't think anyone has argued that this patch is "too risky",
just that "we're trying to stabilize."  Notwithstanding that stability effort,
could we not get this patch checked in?  If not, can we have some specific risk
analysis for why not?

/be

Comment 23

[Marc Attinasi]

The patch does not look risky. It is very localized. I have applied the patch 
and the testcase looks great. Thanks for the patch, timeless! Would you like me 
to check it in, or do you have checkin privs?

Comment 24

[timeless]

I do not have cvs privs, i should get them soon, but i haven't gotten around to 
doing the paperwork. Please commit, thank you.

Comment 25

[Marc Attinasi]

Patch checked in, thanks again timeless!

Comment 26

[Christine Hoffman]

With the following builds:
Win: 7_18_20
Mac: 7_19_12
Linux: 7_20_09
Using the testcase provided, verifying bug fixed.