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 152456 - Better support for partial transparency in indexed images
Better support for partial transparency in indexed images
Status: RESOLVED DUPLICATE of bug 86627
Product: GIMP
Classification: Other
Component: User Interface
2.0.x
Other Linux
: Normal enhancement
: ---
Assigned To: GIMP Bugs
GIMP Bugs
Depends on:
Blocks:
 
 
Reported: 2004-09-12 20:49 UTC by Jim Higson
Modified: 2004-09-13 17:22 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Jim Higson 2004-09-12 20:49:53 UTC
For example, try opening this image in the GIMP: 
 
http://www.masmodels.com/images/port-frame-norm-fs8.png 
 
Although it is indexed, given the current limitation of indexed images only 
containing 1 bit alpha, it would seem to make sense to give the option to open 
it as a 32-bit rgba image. 
 
For example, a dialog with text such as this could be presented when the user 
tried to open such an image: 
 
Image foo.png is indexed, with alpha transparency. Unfortunately this version 
of the GIMP cannot edit indexed images with alpha channels. Do you wish to: 
* Open as an indexed image and loose some transpaency information 
* Open as an rgb image, which can express the image exactly but increases the 
filesize
Comment 1 weskaggs 2004-09-12 22:21:58 UTC
There's a bit of confusion here.  It actually is not really true that indexed
images can only contain 1 bit alpha.  What *is* true is that GIMP *displays*
indexed images as though they contained 1 bit alpha.  It does this because of a
bug in Internet Explorer:  GIMP shows you the image the way it would look if you
viewed it in IE.  GIMP does this in order to protect you from creating images
that look wrong in the most widely used browser.  Pretty much the only reason
people create indexed images is for browsers, so this is important.

What is also true is that if you convert an image from Indexed to RGB mode in
GIMP, your options are either no transparency or 1 bit transparency.  GIMP
enforces this restriction for the same reason as above:  to keep you from
creating things that IE doesn't support properly.

Actually, your image is being loaded in a completely correct way, and if you
convert it to RGB mode after loading, you will see that all the information is
there.  You just don't see it as long as you are in indexed mode.

I know it's strange and confusing.  Blame Microsoft.
Comment 2 Jim Higson 2004-09-12 22:45:06 UTC
This seems absolutely bizare to me - for image editing software to assume it 
knows beter than the user what the images are for. 
 
At the very least this behaviour should be an option. For example a combobox 
labelled "view as would be displayed in..." on the image window. Restricting 
everyone to the lowest common denominator seems wrong. 
 
In my webpages I serve image formats (Gif, PNG32, PNG8) based on content 
negotiation. I am aware of the limitations of IE (and some other, older 
browsers), but don't want my image editing software to double guess what I'm 
trying to do. 
 
Is it the case that I simply cannot hand-tune an 8bit, indexed image in the 
GIMP? 
 
I blame MS for IE not showing PNGs properly, and a bunch of other stuff, but 
not this! 
Comment 3 Sven Neumann 2004-09-13 07:54:30 UTC
Bill, it is actually not Microsoft who is to blame here. The way that indexed
images are handles in GIMP was designed for handling GIF images. At the time
when this decision was made and this code was written, the GIF format was the
format for indexed images with transparency and it only supports 1bit transparency.

Noone really likes how indexed images are handled but so far none of the
developers cares enough to change it. I would actually prefer to see support for
indexed images removed from the GIMP core and let the file plug-ins handle the
conversion.
Comment 4 weskaggs 2004-09-13 15:00:43 UTC
Okay, I stand corrected.  I will reopen this bug, changing the summary and
priority according to the real issue.
Comment 5 Sven Neumann 2004-09-13 15:35:23 UTC
It basically has become a duplicate of bug #86627 then.
Comment 6 weskaggs 2004-09-13 17:22:17 UTC

*** This bug has been marked as a duplicate of 86627 ***