|Summary:||[Windows] Printing of Plugins Ignores Header Settings When Calculating Page Position|
|Product:||WebKit||Reporter:||Brent Fulgham <bfulgham>|
|Component:||WebCore Misc.||Assignee:||Brent Fulgham <bfulgham>|
|Version:||528+ (Nightly build)|
Description Brent Fulgham 2009-12-23 12:35:51 PST
When printing plugins in Windows, the device context handed to the plugin is translated to reflect its position on the page. For example, a widget taking up a region starting at (8, 54) on the screen is printed using a context whose world transform has been translated by (8, 54). This is not correct when a header is being used. In printing tests I can see that the translated context is the same as the on-screen representation, yet the rest of the page contents are further down the page due to the header spacing. This causes the plugin to paint over the top of some of the text region.
Comment 1 Brent Fulgham 2009-12-23 12:45:18 PST
Created attachment 45447 [details] Screen shot of PDF generated by printing a sample page with a single embedded plugin. Notice how the region drawn by the plugin overlaps the heading region (labeled 'Sample Heading'). The region on-screen is properly offset by the amount of the header.
Comment 2 Brent Fulgham 2010-01-05 16:55:16 PST
Created attachment 45941 [details] Patch to update CTM with scale and position when drawing a plugin The attached patch modifies the CTM to account for desired page position, so that margins and header are accounted for when drawing the plugin.
Comment 3 WebKit Review Bot 2010-01-05 17:18:09 PST
style-queue ran check-webkit-style on attachment 45941 [details] without any errors.
Comment 4 Jon Honeycutt 2010-01-07 14:00:48 PST
Comment on attachment 45941 [details] Patch to update CTM with scale and position when drawing a plugin > Index: WebCore/ChangeLog > =================================================================== > --- WebCore/ChangeLog (revision 52752) > +++ WebCore/ChangeLog (working copy) > @@ -1,3 +1,16 @@ > +2010-01-05 Brent Fulgham <firstname.lastname@example.org> > + > + Reviewed by NOBODY (OOPS!). > + > + Scale device context world transform passed to plugins to account > + for print surface DPI. > + http://bugs.webkit.org/show_bug.cgi?id=33022. This mentions a different bug #. Is there a similar change that needs to be made for windowless plug-ins? Is there any way to add a test this? Perhaps a manual test? r=me