After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 91941 - Layer Move Independent of Layer Mask
Layer Move Independent of Layer Mask
Status: RESOLVED OBSOLETE
Product: GIMP
Classification: Other
Component: User Interface
1.x
Other All
: Normal enhancement
: Future
Assigned To: GIMP Bugs
GIMP Bugs
: 131163 136649 313395 (view as bug list)
Depends on: 93639
Blocks:
 
 
Reported: 2002-08-29 03:18 UTC by lasm
Modified: 2018-05-24 10:46 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Mockup of a mask lock. The upper layer with mask will move the mask independently from the layer. The lower layer with mask has them locked, so they will both move/transform. (8.20 KB, image/png)
2003-05-29 14:46 UTC, Jakub Steiner
Details

Description lasm 2002-08-29 03:18:23 UTC
Currently there is a little "bug" with moving the layer.
If the layer has a layer mask, moving the layer, moves the mask as well.

See the picture http://www.geocities.com/lasm.rm/layer-bug1.jpg

This is what is shown in the layer dialog
http://www.geocities.com/lasm.rm/layer-bug2.jpg

There is a little refresh problem here as the layer thumbnail should be
be moved to the left and upwards. 

The desired enhancement is shown in this picture
http://www.geocities.com/lasm.rm/layer-bug-desired1.jpg

Here moving the layer DOES NOT automatically move the layer mask along 
with it.

To accomplish this enhanced feature, it is suggested that there should be
a little chain icon in between the layer thumbnail and the layer-mask 
thumbnail as shown in this composited picture
http://www.geocities.com/lasm.rm/layer-bug-desired2.jpg

When the chain is broken, the layer mask does not move, only the layer 
moves.

When the chain is toggled to "unbroken", then the layer mask moves along
with the layer, as in current "buggy" implementation.

Note that this little enhancement will open up a whole new world of 
possibilities for the user, and it is in fact a standard feature for years 
in commercial packages such as Paint Shop Pro and Photoshop.
Comment 1 Raphaël Quinet 2002-08-29 16:35:44 UTC
I do not consider the current behavior to be wrong.  On the contrary,
this is the most intuitive from my point of view: the mask is a part
of the layer so moving the layer should move the mask as well.  But
I agree that in some rare cases it could be useful to be able to move
one of them separately from the other one.  So I will confirm this
enhancement proposal.

Note that all of your images are broken.  I suspect that geocities
does not allow external referers to images, only to top-level HTML
pages so that they can display their advertisements.  It would be
much better if you could attach the images to this bug report so that
we can see them easily (even if geocities or your pages disapear in
the future).
Comment 2 lasm 2002-08-30 03:48:55 UTC
I am sorry I wasn't aware that the geocities pictures link are broken
without logging in. Putting them in a html page seems to work however,
so here it is again

http://www.geocities.com/lasm.rm/layer-bug91941.html

BTW, I haven't seen anyone attach pictures to bug reports, is that
possible ?
Comment 3 Jakub Steiner 2002-08-30 13:15:50 UTC
I strongly vote for this to be implemented. The only way to move the
mask independently (if I'm not mistaken) is to select all (ctrl+A) and
float the mask (ctrl+shift+L) and then move. It gets problematic if
the mask is larger than the layer, since ctrl+A only selects the image
size area.

I'd love to see that link icon between the layer and mask, but make
that on by default, so that moving a layer moves a mask too. But the
user is free to move them separately.
Comment 4 Raphaël Quinet 2002-08-30 15:43:11 UTC
lasm: You can easily attach images to the bug reports.  If you look
around, you will see that several GIMP bug reports already have test
images attached to them (among the recent ones, you can look at bug
#79023, bug #84640, bug #86290, ...).  To attach your image, just
follow the link "Create a new attachment" and select the correct type
for the image that you want to upload.  This would be more convenient
than having to go through geocities and get all their advertisements
before the interesting images.

Jakub: I switched the priority to "High".  This will not change much
until someone volunteers to implement that, but at least this is now
marked as "High priority".

Just out of curiosity, how did you get a mask that is larger than
the layer?
Comment 5 Dave Neary 2002-08-30 16:15:19 UTC
> Just out of curiosity, how did you get a mask that is larger than
> the layer?

For example, create a text layer, and create another layer the same
size as the image, in which you put a gradient. On the text layer, add
a layer mask, copy the gradient layer, and paste it into the mask. The
mask is teh same size as the image, the layer is big enough to hold
the text. Et voilá.

Dave.
Comment 6 Jakub Steiner 2003-05-29 14:46:35 UTC
Created attachment 16956 [details]
Mockup of a mask lock. The upper layer with mask will move the mask independently from the layer. The lower layer with mask has them locked, so they will both move/transform.
Comment 7 Dave Neary 2003-07-26 20:59:51 UTC
Changing to future, it seems unrealistic for this to be done before camp.

Dave.
Comment 8 Michael Natterer 2004-01-21 13:03:59 UTC
For the reasons given in bug #131163, I vote for closing
this one as WONTFIX or NOTABUG.
Comment 9 Michael Natterer 2004-01-22 16:08:59 UTC
*** Bug 131163 has been marked as a duplicate of this bug. ***
Comment 10 lasm 2004-02-18 05:17:03 UTC
Layer Mask Movement Semantics

Someone asked what should be the semantics for layer mask movement in 
bug 131163 which I summarize briefly as follows.

A) Layer Mask Linked to Layer
When the layer mask is linked to layer, this is the simplest case.
There is no change here. The layer will move together with the layer 
mask, and whatever area that is exposed will be filled with alpha 
transparency.

B) Layer Mask Un-linked to Layer (Independent Move)
For this case, the layer mask is free to move independently of the 
layer itself. When the layer mask exposes new areas, it should be 
filled with black color (if layer mask was created with Black-full 
transparency OR created with Layer's Alpha Channel)

The exposed region should be filled with white only if the layer mask 
was created with White (full opacity)

When the exposed layers are later covered, the previously stored mask 
in the buffer should be used to restore the mask.

I think there should be a 3x3 matrix for buffering the layer mask, 
with the centre of the matrix being the currently visible layer mask.

If the layer mask is moved more than twice the length of the mask 
horizontally, the protruding part should be clipped.

Similarly if the layer mask is moved more than twice the height of 
the mask vertically, the protruding region should be clipped and lost.

Hopefully this semantics is clear, and as far as I can tell, it is 
pretty consistent with other programs, including PS.
Comment 11 Raphaël Quinet 2004-02-19 12:13:44 UTC
I suggest marking this bug as depending on bug #93639 (which should
allow layers to grow automatically) because growing the mask while it
is moved will require some changes in the tile management that are very
similar to the ones discussed in bug #93639.  Look at my comment from
2003-12-12 and the following one by Daniel if you are confused by some
of the earlier comments in that bug report.
Comment 12 Sven Neumann 2004-03-09 15:30:41 UTC
*** Bug 136649 has been marked as a duplicate of this bug. ***
Comment 13 Michael Natterer 2005-08-13 15:02:57 UTC
*** Bug 313395 has been marked as a duplicate of this bug. ***
Comment 14 Martin Nordholts 2010-06-29 18:58:33 UTC
Closing at WONTFIX. With automatically managed layer boundaries, there is no need to add any dedicated functionality or code for this.
Comment 15 Joao S. O. Bueno 2017-06-22 08:57:05 UTC
So - the feature request closed 7 years ago is reopened. While "automatically managed layer boundaries" did not come by (yet), nor do I see how that could provide the features of moving the layer contents separated from the mask, it would be nice if the parties interested in resurrecting this would throw in some more context.
Comment 16 GNOME Infrastructure Team 2018-05-24 10:46:12 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to GNOME's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/gimp/issues/31.