You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have come to this via a few steps (un rotated image in a weasyprint PDF, through image metadata, through GDK-Pixbuf, through weasyprint source code, to here) so I have not had any involvement in this project and I also have very limited experience with the CFFI.
What I think I have found is that when calling GDK-Pixbuf the gtk.gdk.Pixbuf.apply_embedded_orientation function isn't being called so images with embedded orientation information (in my instance iPhone photos) are not having that information read and applied. In my case this is resulting in portrait orientation images displaying in PDF's as landscape. So far I am not convinced that this is a problem (I am setting height for these images to ensure I fit two stacked on an A4 page) so this way around makes best sense to me for my specific use case.
This behaviour differs from what I see loading the same image directly into google chrome which would often seem to be the default expectation by end users. However this not being the default any change may be more significant as people may have made assumptions / changes on their own that could conflict quite interestingly.
So I wanted to ask a question, considering the minimum required GDK-Pixbuf version already includes this "apply embedded orientation" function has this been considered as an option that should be used by default? or should be available as an option at calling time for users to knowingly switch to this if the current default is not their preference?
As I said I have limited experience in lower level coding and so any such pull request, even if I got it ready, would be expected to be rough compared to other code quality or calling it the "best" way.
Regards
Alexander
The text was updated successfully, but these errors were encountered:
I have come to this via a few steps (un rotated image in a weasyprint PDF, through image metadata, through GDK-Pixbuf, through weasyprint source code, to here) so I have not had any involvement in this project and I also have very limited experience with the CFFI.
What I think I have found is that when calling GDK-Pixbuf the gtk.gdk.Pixbuf.apply_embedded_orientation function isn't being called so images with embedded orientation information (in my instance iPhone photos) are not having that information read and applied. In my case this is resulting in portrait orientation images displaying in PDF's as landscape. So far I am not convinced that this is a problem (I am setting height for these images to ensure I fit two stacked on an A4 page) so this way around makes best sense to me for my specific use case.
This behaviour differs from what I see loading the same image directly into google chrome which would often seem to be the default expectation by end users. However this not being the default any change may be more significant as people may have made assumptions / changes on their own that could conflict quite interestingly.
So I wanted to ask a question, considering the minimum required GDK-Pixbuf version already includes this "apply embedded orientation" function has this been considered as an option that should be used by default? or should be available as an option at calling time for users to knowingly switch to this if the current default is not their preference?
As I said I have limited experience in lower level coding and so any such pull request, even if I got it ready, would be expected to be rough compared to other code quality or calling it the "best" way.
Regards
Alexander
The text was updated successfully, but these errors were encountered: