1515 - for dotted borders, want round dots instead of square dots

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, enhancement, P2)

Description

[howcome]

The CSS specification do not detail how dotted borders should be drawn. Opera
3.50 draws them as "zebra crossings" when the border width increases. I think
that is a better solution than nglayout's current "square dots" policy. Section
(5) of the test page lets you test this.

Comment 1

[Peter Linss]

How about actual round dots for dotted borders?

Comment 2

[Hixie (not reading bugmail)]

Now that would be cool.

Comment 3

[leger]

Moving to M11. Not an M10 blocker.

Comment 4

[dcone (gone)]

This won't be addressed in the near future, maybee in the next version... or if
there is time...

Comment 5

[Christine Hoffman]

Verified  REMIND

Comment 6

[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 7

[Matthew Brealey]

This has a target of M11. That was some time ago.

Comment 8

[Hixie (not reading bugmail)]

Reopening and moving to future...

Comment 9

[ekrock's old account (dead)]

Nominating nsbeta2,nsbeta3,rtm. Recc. nsbeta2+, falling through to nsbeta3+ and 
rtm+ if necessary. This is a W3C CSS1 Official Test Suite compliance bug.

Comment 10

[ekrock's old account (dead)]

Sorry, to clarify: this is not an nsbeta2 stopper, but I'd permit checking in 
prior to nsbeta2 if we have a fix. Recc. nsbeta2+[some lenient date-], falling 
through to nsbeta3+ hard stop if not fixed during nsbeta2. This is a W3C CSS1 
Official Test Suite compliance bug.

Comment 11

[Hixie (not reading bugmail)]

Eric: This is possibly one of the bugs I would delay the most to fix, and is
IMHO not even remotely important for FCS. There is no _bug_ here, we are merely
suggested that the dots of 'dotted' borders be drawn as round dots instead of
square ones.

Are you sure this should be nsbeta3 hardstop?

Comment 12

[leger]

Putting on [nsbeta2-] radar. Not critical to beta2. 

Comment 13

[ekrock's old account (dead)]

Sorry, and thanks Ian! I totally misunderstood this report; I thought we were 
*failing* the test. Ian's correct; there's not bug & this goes to future.

Clarifying Summary from "dotted borders" to "for dotted borders, want round dots 
instead of square dots." 

To simplify tracking & searching (in particular, to avoid wasting the time of 
people who will be reviewing the [nsbeta2-] bugs for nsbeta3, etc., I'm taking 
the unusual (but in this case reasonable) step of removing all of the 
PDT-related markings from both Keywords and Status Whiteboard. (It was only my 
goof that put them there in the first place.) Removing "nsbeta2,nsbeta3,rtm" 
from Keywords and "[nsbeta2-]" from Status Whiteboard.

Also Ian, should this really have the "css1" keyword? It's not a compliance 
issue. Your call.

Comment 14

[Hixie (not reading bugmail)]

Eric: I'm using the 'css1' keyword for anything CSS1-related, not just for 
compliance issues. Almost all the bugs marked CSS1 which are currently Future,
for example, are not actually compliance issues.

Comment 15

[Hixie (not reading bugmail)]

Internet Explorer 5.5 already does this, for the record :-)

Comment 16

[dcone (gone)]

I am marking this as invalid for a few reasons.  
1.)  Currently all our border code goes thru routines that are set up to draw 
rectangular type dots, so dashes and dotted use the same code as well as all 
border styles.  To change this would be a pretty big job.. not worth the errors 
that might be introduced.
2.) Using dots could complicate some issues like going around corners.
3.) Looking at examples in books.. the way we implemented dotted and dashed seem 
consistent with how they handle this issue.  
4.) This is so low a priority that even if we did decide to do it.. it would 
probably never be addressed..

Comment 17

[Hixie (not reading bugmail)]

Reopening. INVALID is the wrong resolution, it should be WONTFIX if we actually
did decide to close it unfixed.

However the competition successfully implements this -- and so it is a valid 
enhancement request even if it is not a high priority one.

Comment 18

[Hixie (not reading bugmail)]

Moving off dcone's plate and onto the helpwanted pile...

Comment 19

[Hixie (not reading bugmail)]

Netscape's standard compliance QA team reorganised itself once again, so taking 
remaining non-tables style bugs. Sorry about the spam. I tried to get this done 
directly at the database level, but apparently that is "not easy because of the 
shadow db", "plus it screws up the audit trail", so no can do...

Comment 20

[Greg K.]

Dotted borders have square dots on Mac, too. Changing to All/All.

Comment 21

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Taking (and ->P2), although I may not get to it anytime soon.

Comment 22

[Brant Gurganus]

Just an FYI: The Oxford American Dictionary says a dot is "a small, *round* point".

But, in my opinion, what are the advantages to having round dots for dotted
borders anyway?  A dotted border is under most, if not all, circumstances, so
small that you can't tell if it is a round dot or a rectangular dot anyway.

Comment 23

[Eric A. Meyer (dead account)]

In most cases, yes, but not all.  If you're into CSS semi-hacks, it's possible
in IE5/Mac to do this:

   <span style="border-left: 1em dotted purple;">blah</span>

...and get a single purple circle to the left of the content of the element. 
For that matter, defining a truly huge single-dot border can give you non-image,
completely text-zoom-scaling circles.

Granted, it's an edge case, but it's still something Gecko still can't do.  And
even in less extreme cases, an author might want a 5px wide border with round
dots in it instead of smaller dashes.  So we really should, at some point, get
our dots to be actual dots.  I wouldn't call it a top priority or anything but
it would certainly be nice.

Comment 24

[Grey Nicholson (they/them)]

Every definition I can find for "dot" says dots are round, and I agree -
besides, it looks nicer.

Comment 25

[Peter Kasting]

Taking for preliminary investigation.  biesi suggests looking for this code in nsCSSRendering.cpp.

Comment 26

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

See also bug 19963, which requires more significant changes to dotted and dashed borders, and is actually somewhat necessary for colors with alpha, which we now support with cairo builds.

Comment 27

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Note that one of the interesting questions is what to do at the corners when dotted borders have different widths.  I suggest an ellipse, since that allows the two sides of the diagonal to be drawn independently, which does the right thing in the normal case (same width) and does something at least somewhat reasonable in cases where one side is dotted an the other is not.  But you might want to test what other browsers do there as well (but consider their solutions in light of how well they work for bug 19963 and how well they work for rgba colors).

Comment 28

[Peter Kasting]

This would be a nice nhancement, but I'm not going to get it in the foreseeable future, so I'm going to go ahead and dump it to nobody.

Comment 29

[Pam Greene]

A proposal for the end result of dots and bevels has been attached to bug 19963.

Comment 30

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Fixed by thebes rendering rewrite; dots are used for dotted borders > 2px in size.

Comment 31

[Stefan Mayrhofer]

Attachment: screenshot of the new round-dotted border

Maybe it is just me, but personally I would not call this fixed.
The circles are cut off at the corners.

Comment 32

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Fixing that is bug 379303 -- should be straightforward, just haven't written the code yet.