Confirmed.  It is easiest to see this bug by slowly going over a Wikipedia icon from 
all directions and seeing where the open hand changes to the pointing hand.  Compare 
to when it changes in Firefox.

(No comment was entered for this change.)

Taking these to triage.

You should double-check me but I believe this is fixed with the nightly builds of 
Chromium.  You can get them from http://build.chromium.org/buildbot/continuous/LATEST/

The finger appears where the tip of the mouse cursor was.  This is correct, I 
believe.

Unless I hear otherwise I will close this as fixed on December 1st.

Which hand cursor are you guys checking?  The pointing-hand cursor placement is 
correct.  It is the full, open-hand cursor that we're talking about here.

It is still showing the hand being different compared to what Firefox and IE have.  
Both FF & IE have the tip of the arrow pointer = pretty close to the base of the 
joint between index and middle finger on the hand cursor.

Safari, on the other hand, does exactly what Chrome/ium is doing, placing the active 
spot very close to the upper left 0,0 loc of the hand cursor.

Did you guys try loading up Google maps, turn on the Wikipedia feature (under More), 
and slowly drag the cursor onto the W icon from left, right, top and bottom?  Compare 
how FF & IE do it.  The FF & IE placement are more natural than the Safari & Chrome 
placement.

Oh, and for an even clearer example, use the tiny W's instead of the large ones.  
This will clearly show you why this is a problem.  Other browsers you can click those 
tiny W's very easy, very naturally.

The additional examples make it clearer what should be fixed.  This is a good candidate for someone who wants 
to learn the Chromium codebase and make their first patch.  

(No comment was entered for this change.)

Sorry, this should be upstreamed to WebKit as it also happens in Safari.

I tried this in build 2.0.160.1 (Developer Build 9120), It is almost same in Chrome 
and IE, and I feel the hot-spot is place properly.

mfosteru..., can you verify this in the latest build?

Looks like this is fixed in the current release (1.0.154.48). Cheers.

Yup, looks good here also in 1.0.154.48.  Keep up the great work guys!

Close it then.

Don't close it. The fix seems to only work for Google Maps. If I use the exact same 
cursor file in a test page, the hotspot goes back to broken. (I have 1.0.154.48 on 
Windows.) Attaching a test page.

	cursor-test.html
	192 bytes Download

Please try this against a 2.x build.

vladimir.slepnev, what is broken? 

I tried you cursor-test.html, "hand" cursor is shown when mouse on any of body area,
which is same with IE (in Firefox, "hand" cursor is shown only when mouse on "Hello
World!" text), this should not be a problem.

Actually there is no hotspot in your test case...

mdu, the hotspot is specified inside the .cur file that's taken directly from Google 
Maps (read the html). To verify the placement, move the mouse between the document 
body and Chrome's top bar - when the cursor switches from arrow to 
"openhand_8_8.cur", the hand is off to the right. Compare with Google Maps when you 
enter or leave the map area.

The .cur file is just an image, there is no hotspot in the case, but when moving
mouse between document body and chrome's top bar, it truly has a little difference
between Chrome/Safari and IE/Firefox.

Supposing moving mouse from chrome's top bar down to body area:

In Chrome/Safari, "hand" cursor is changed to "pointer", the pointer is put on the
location of "middle finger" of the previous "hand" cursor.

In IE/Firefox, "hand" cursor is changed to "pointer", the pointer is put on the
location of "fore finger" of the previous "hand" cursor.

mdu,

1) The test case doesn't specify a hotspot in CSS, because Google Maps doesn't do it 
either (I checked). The .cur file has a hotspot specified _inside_ it. Look up the 
.cur format on Wikipedia. Where do you think Firefox and IE get the hotspot position? This particular cursor has size 32x32 and hotspot at 8,8, which I learned by using a 
special .cur file editor.

2) The vertical position of the hand is also off. Move the mouse really slowly. In 
IE/Firefox the palm of the hand appears over the tip of the mouse pointer. In 
Chrome/Safari it jumps down - the hotspot seems to default to 0,0. I'm mystified why 
this doesn't happen on Google Maps.

Thanks for this dicussion thread. I had to add a background gif to get ie and firefox
to recognize the same hotspot. ff was recognizing the entire div, but ie was only
responding to the text/font area. thus, tree menu was enabled.

(No comment was entered for this change.)

(No comment was entered for this change.)

Moving out of Mstone:2.1 as there just isn't enough time to work on this 
issue.

(No comment was entered for this change.)

(No comment was entered for this change.)

(No comment was entered for this change.)

What is the current state of this bug?  Still looks broken to us when an application
specifies a .cur file using a CSS rule for an element.

We use a custom cursor which has the hot spot in the bottom right corner. IE8 and FF3
get it right. Chrome and Safari ignore this data and use the top left as the hot
spot. As noted by Jon in Comment 9, this would appear to be a WebKit issue [To
complete the list, Opera9 ignores the CSS rule altogether and we see the default hand]

I'd like to add that this bug affects not only websites using custom cursors, but also 
Chrome itself.
When you press middle mouse button to activate auto scrolling, a four-arrow cursor in a 
circle appears. It has hot spot in the center, as expected. However, once you move the 
mouse to start scrolling, cursor changes to directional arrow that is offset to the 
right and downward, i.e. its hot spot is in the top left corner instead of the center.

looks like this might be related to: https://bugs.webkit.org/show_bug.cgi?id=15779 
having said that, it's been sitting there for a long time. 

FYI, google fixes this issue on Maps with a CSS3 custom cursor hot-spot declaration.

It's described toward the middle of:
https://developer.mozilla.org/en/Using_URL_values_for_the_cursor_property

Their full CSS is probably cursor: url(http://url/to/openhand_8_8.cur 8 8, default;
which moves the hotspot to the center of the hand.  The "8 8" doesn't show up in Web 
Inspector which makes it hard to find.

This doesn't fix this bug, which is probably a Webkit bug because it also happens in 
Safari, but it is a possible workaround until Webkit fixes it...

Apologies for the typo, I didn't close my parens:
cursor: url(http://url/to/openhand_8_8.cur 8 8, default;

Odd. My close-paren gets munged.  Well, you can go to the mozilla page for the correct 
declaration.

This looks like webkit bug https://bugs.webkit.org/show_bug.cgi?id=8989, as far as I 
can tell.
This is not fixed in the latest webkit nightly (r51062)

 Issue 12196  has been merged into this issue.