Cursor glyph for CSS3 'all-scroll' is the same as for 'move'.
(spawned off from bug 163174 comment 98)
Bug 163174 comment 98 suggests using the autoscroll icon/cursor now in use
in Firefox. Another alternative is to use the same glyph as IE6.
Gecko-based K-meleon 0.9b uses all same 9 all-scroll cursors as in/with MSIE 6;
so maybe there is a way to avoid duplication of work here.
If the all-scroll cursor we intend to implement does not use/reuse the
auto-scroll image in Firefox (and installable in Mozilla as an extension), then
I don't think we're making the best out of the situation. In any case, I can
create all 9 all-scroll cursors in win32 .cur files without a problem, in less
than an hour or so. Personally, I think the auto-scroll image in Firefox should
be a real cursor, not an image.
You can view the win32 .cur
in/with MSIE 6 at
Just mouse over the all-scroll table row. As I said, I can do all 8 others very
fast if needed. I'll do one for bug 258960 today...
Great work! However, your cursor is invisible when used above dark backgrounds,
especially on black. And it lacks a hotspot.
Use bug 275173 comment 5 and 6, or let me fix it.
Created attachment 169060 [details]
Microsoft Windows custom cursor (.CUR) for CSS3 all-scroll
1) Hotspot defined in the center of the glyph
2) Overall cursor size reduced to 326 bytes instead of 2238
3) Edge pixels transformed from black to reversed background, to make the
cursor visible over black-coloured webpages
I am not sure about the 3rd idea -- whether it looks good or not. What were
other variants? With black pixels untouched and white halo added, the glyph
looks ugly on dark blue background. If all pixels are either transparent or
reverse, then the whole thing is invisible on 50% gray. I've tried to reach
> Advancements made:
> 1) Hotspot defined in the center of the glyph
I defined it at the center+top of the glyph. I don't know what's the spec for this.
> 2) Overall cursor size reduced to 326 bytes instead of 2238
I just re-created one with only 2 colors (black and white) and got 326 bytes as
well.... which makes much more sense. (I need to make such kinds of adjustements
in bug 258960#c3. The cursor glyph is bigger (and as big as the one in MSIE 6)
than yours in attachment 169060 [details].
> 3) Edge pixels transformed from black to reversed background, to make the
> cursor visible over black-coloured webpages
I do not know how you transform to reverse background... still exploring
MicroAngelo on this.
> I am not sure about the 3rd idea -- whether it looks good or not. What were
> other variants? With black pixels untouched and white halo added, the glyph
> looks ugly on dark blue background. If all pixels are either transparent or
> reverse, then the whole thing is invisible on 50% gray. I've tried to reach
> some compromise.
You can view my latest attempt at all-scroll win32 .cur
in/with MSIE 6 at
When I have time, I will create a more versatile testing div so that the user
could select a .cur by himself and even choose color background with a color picker.
> > 1) Hotspot defined in the center of the glyph
> I defined it at the center+top of the glyph. I don't know what's the spec for
I checked with MSIE 6 and empirically was able to locate the hot spot in the
center of the glyph.
Created attachment 169214 [details]
Custom made CSS3 all-scroll cursor (.cur)
This one does not use background reverse color. Shape, size and outline match
exactly what MSIE 6 does.
Removing helpwanted and qawanted keywords.
If we agree on the cursor, then the only thing remaining is editing files
mentioned in bug 258960#c2. As far as I'm concerned, this bug is almost/pretty
much fixed. The "art" work has been done.
Same thing with other CSS3 cursor bugfiles: bug 275173, bug 258960.
Finally, related to this bugfile, it would be more logical if auto-scroll image
in Firefox (and as an Mozilla extension) would use the all-scroll cursor
instead: we would be conveniently closer to MSIE 6. Best, ideal would be to do
what K-meleon 0.9b does.
Reinstalling helpwanted keyword since someone will have to implement this.
I'll do the 8 direction-cursors while using the images at
This bug is somewhat related to bug 304563