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 23610 - Resizing seems to work but the saved image still has old dimensions.
Resizing seems to work but the saved image still has old dimensions.
Status: VERIFIED DUPLICATE of bug 51114
Product: GIMP
Classification: Other
Component: General
1.x
Other Linux
: Normal normal
: ---
Assigned To: Daniel Egger
Daniel Egger
Depends on:
Blocks:
 
 
Reported: 2000-09-07 12:00 UTC by carljohan
Modified: 2009-08-15 18:40 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description carljohan 2001-01-28 16:12:01 UTC
Package: gimp
Version: 1.0.4

Name........: Carl-Johan Kjellander
Email.......: carljohan@kjellander.com
Platform....: Dual Celeron PC, Linux 2.2.x
GIMP Version: 1.0.4
GTK Version.: RedHat 6.1. don't know which.
WM/Version..: Window Maker 0.61.0


-- Other system notes:

--

-- Problem description:
I have a png image which is 800x654. I resize it to 800x600
and save the file. when I open it again (in gimp or any
other program) it is still 800x654.
--


-- How to repeat:
Open a picture in gimp. Resize it. Save it. Open it again.
Still the old size.
--


-- Other comments:

--




------- Additional Comments From quinet@gamers.org 2000-09-21 11:29:12 ----

Subject: Re: [gimp-bug] Load, Crop, then Save results in original (not cropped) image being saved
From: quinet@gamers.org (Raphael Quinet)
To: 23610@bugs.gnome.org, 25266@bugs.gnome.org
Message-Id: <200009211529.RAA12119@chapelle.eed.ericsson.se>
Date: Thu, 21 Sep 2000 17:29:12 +0200 (MET DST)

On Wed, 20 Sep 2000, Austin Donnelly <austin@gimp.org> wrote:
> On Wednesday, 20 Sep 2000, quinet@gamers.org wrote:
> > If you load an image (containing one or several layers) and
> > the first operation that you do on this image is to crop a
> > part of it with the crop tool, then the canvas is resized
> > but the layers do not seem to be affected.  If you then save
> > the image without flattening it first, then the original
> > layers are saved, ignoring the area that should have been
> > cropped.
> 
> This bug (or one very similar to it) is also present in gimp 1.0.4,
> and already has a bug open on it: see bug report #23610.  I can
> reproduce that bug on 1.0.4 and 1.1.25, so I expect yours will fall
> into the same category.

I don't think that the two bugs are really related.

Bug #25266 is related to a strange behaviour of the crop tool that
only happens if cropping is the very first operation that you perform
on an image.  I cannot reproduce the bug if I do anything on the image
before cropping it, or if I create a new image instead of loading one
from disk (and even then, it depends on how you load the image).  See
the bug report for other strange details about that bug.  It smells
very much like uninitialized memory or something like that.

Bug #23610, on the other hand, happens in a consistent way and is
related to a design problem in the File->Save plug-ins.  There is
nothing random about it, and the problem can be explained very easily
(if I am not mistaken).  The problem is that the file plug-ins that
only support a single layer will not trigger the "Export" feature when
they are told to save an image that contains only one layer.  The
plug-in uses only the current drawable (layer) and ignores the canvas
size.  Most file plug-ins (except for PNG) ignore the layer offset as
well.  As a result, what gets saved in the file is the whole layer,
even if only a part of it should be visible through the canvas.
This problem is also described in bug #25272: the thumbnails are
generated from the visible area, not from what gets saved, so the
result is even more confusing.

So the real problem of bug #23610 is that the File->Save plug-ins
should compare the canvas size with the layer size and also check if
the layer offset is <0,0> before attempting to save the layer.  If
they do not match, then only the visible part of the layer should be
saved.  Another solution (at least for Gimp 1.1.x) would be to trigger
the "Export" option when the Gimp detects a mismatch in the layer size
or offet.  Flattening the image as part of the export process would
solve the problem described here.

The results are interesting with a PNG image, because the plug-in (in
version 1.1.25) checks the layer offsets but ignores the canvas size.
Try this for example: create or load an image in which all parts of
the image can be easily identified (so that you can see if any offset
is applied).  Then resize the image to make it smaller, and move the
layer so that you only see some area near the middle of the original
image.  Now save it as PNG, and load the result.  You will get an
image that has the original size because the layer size was used
instead of the canvas size when saving, but parts of the layer will be
out of the visible area because the offsets were saved in the PNG
file.  The other file plug-ins ignore both the canvas size and the
offset, so it looks like no resizing occured.

Hmmm...  Now that I think about it, there is also a problem if you set
the opacity of the layer to something less than 100%.  What you see is
a partially transparent layer, but what is saved is the fully opaque
layer.  This is another aspect of the problem of saving a layer as if
it was isolated instead of considering all attributes of the image.

So IMHO the bugs #23610, #25266 and #25272 have three separate causes
even if they seem to be related:
- #23610 is a problem in the file plug-ins.  A workaround would be to
  use the Export option even on single-layer images, if the canvas
  size does not match the layer size or if the offets are non-zero.
- #25266 seems to be a problem with the crop tool, although the exact
  cause is still a mystery to me.
- #25272 is a problem in the way the thumbnails are generated.  It is
  affected by #23610, but can also be caused by other means (e.g. if
  a channel mask is active).

-Raphael




------- Additional Comments From quinet@gamers.org 2000-09-21 11:29:12 ----

Subject: Re: [gimp-bug] Load, Crop, then Save results in original (not cropped) image being saved
From: quinet@gamers.org (Raphael Quinet)
To: 23610@bugs.gnome.org, 25266@bugs.gnome.org
Message-Id: <200009211529.RAA12119@chapelle.eed.ericsson.se>
Date: Thu, 21 Sep 2000 17:29:12 +0200 (MET DST)

On Wed, 20 Sep 2000, Austin Donnelly <austin@gimp.org> wrote:
> On Wednesday, 20 Sep 2000, quinet@gamers.org wrote:
> > If you load an image (containing one or several layers) and
> > the first operation that you do on this image is to crop a
> > part of it with the crop tool, then the canvas is resized
> > but the layers do not seem to be affected.  If you then save
> > the image without flattening it first, then the original
> > layers are saved, ignoring the area that should have been
> > cropped.
> 
> This bug (or one very similar to it) is also present in gimp 1.0.4,
> and already has a bug open on it: see bug report #23610.  I can
> reproduce that bug on 1.0.4 and 1.1.25, so I expect yours will fall
> into the same category.

I don't think that the two bugs are really related.

Bug #25266 is related to a strange behaviour of the crop tool that
only happens if cropping is the very first operation that you perform
on an image.  I cannot reproduce the bug if I do anything on the image
before cropping it, or if I create a new image instead of loading one
from disk (and even then, it depends on how you load the image).  See
the bug report for other strange details about that bug.  It smells
very much like uninitialized memory or something like that.

Bug #23610, on the other hand, happens in a consistent way and is
related to a design problem in the File->Save plug-ins.  There is
nothing random about it, and the problem can be explained very easily
(if I am not mistaken).  The problem is that the file plug-ins that
only support a single layer will not trigger the "Export" feature when
they are told to save an image that contains only one layer.  The
plug-in uses only the current drawable (layer) and ignores the canvas
size.  Most file plug-ins (except for PNG) ignore the layer offset as
well.  As a result, what gets saved in the file is the whole layer,
even if only a part of it should be visible through the canvas.
This problem is also described in bug #25272: the thumbnails are
generated from the visible area, not from what gets saved, so the
result is even more confusing.

So the real problem of bug #23610 is that the File->Save plug-ins
should compare the canvas size with the layer size and also check if
the layer offset is <0,0> before attempting to save the layer.  If
they do not match, then only the visible part of the layer should be
saved.  Another solution (at least for Gimp 1.1.x) would be to trigger
the "Export" option when the Gimp detects a mismatch in the layer size
or offet.  Flattening the image as part of the export process would
solve the problem described here.

The results are interesting with a PNG image, because the plug-in (in
version 1.1.25) checks the layer offsets but ignores the canvas size.
Try this for example: create or load an image in which all parts of
the image can be easily identified (so that you can see if any offset
is applied).  Then resize the image to make it smaller, and move the
layer so that you only see some area near the middle of the original
image.  Now save it as PNG, and load the result.  You will get an
image that has the original size because the layer size was used
instead of the canvas size when saving, but parts of the layer will be
out of the visible area because the offsets were saved in the PNG
file.  The other file plug-ins ignore both the canvas size and the
offset, so it looks like no resizing occured.

Hmmm...  Now that I think about it, there is also a problem if you set
the opacity of the layer to something less than 100%.  What you see is
a partially transparent layer, but what is saved is the fully opaque
layer.  This is another aspect of the problem of saving a layer as if
it was isolated instead of considering all attributes of the image.

So IMHO the bugs #23610, #25266 and #25272 have three separate causes
even if they seem to be related:
- #23610 is a problem in the file plug-ins.  A workaround would be to
  use the Export option even on single-layer images, if the canvas
  size does not match the layer size or if the offets are non-zero.
- #25266 seems to be a problem with the crop tool, although the exact
  cause is still a mystery to me.
- #25272 is a problem in the way the thumbnails are generated.  It is
  affected by #23610, but can also be caused by other means (e.g. if
  a channel mask is active).

-Raphael




------- Bug moved to this database by debbugs-export@bugzilla.gnome.org 2001-01-28 11:12 -------
This bug was previously known as bug 23610 at http://bugs.gnome.org/
http://bugs.gnome.org/show_bug.cgi?id=23610
Originally filed under the gimp product and general component.

Unknown version 1.0 in product GIMP. Setting version to the default, "unspecified".
The original reporter (carljohan@kjellander.com) of this bug does not have an account here.
Reassigning to the exporter, debbugs-export@bugzilla.gnome.org.
Reassigning to the default owner of the component, egger@suse.de.

Comment 1 Daniel Egger 2001-02-08 17:30:29 UTC
Try again with GIMP 1.2.1 and file a new bugreport if necessary
(shouldn't though).
Comment 2 Raphaël Quinet 2001-02-19 17:37:45 UTC
This bug is still open in the current version of the Gimp.  I wrote
a more detailled bug report that explains the problem a bit better,
so I am marking this old bug as a duplicate of the new one (#51114).
Comment 3 Raphaël Quinet 2001-02-19 17:38:09 UTC

*** This bug has been marked as a duplicate of 51114 ***
Comment 4 carljohan 2001-02-20 15:45:11 UTC
I tried to repeat the bug with gimp-1.2.1.

-open a 800x654 image
-resize the canvas to 800x600
-save the image.

When saving a window popped up and informed that the image would
be cropped. The resulting image was 800x600 as expected.

Good work. Keep working on #51114.
Comment 5 Raphaël Quinet 2003-06-20 20:09:46 UTC
The fix for bug #51114 was part of the stable release 1.2.3.  Closing
this bug.