| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
|
|
Don't call mkdir explicitly in the build rule. Use the ALL_DIR variable.
|
|
|
|
|
|
This fixes bug 694755. Thank you to Michael Cadihac for the patch
|
|
Zero and one-point case both lead to an empty rectangle, but the one-point case
needs expanding but wasn't because fz_expand_rect treats an empty rectangle
as a special case (as it should)
|
|
Add a function to clone stroke states, a magic number to keep in the
reference count to signal that a stroke state is stack-stored, and
automatically clone stack stored stroke states in the keep function.
Use fz_default_stroke_state to initialise stack stored stroke states.
|
|
Self balancing AA-tree.
|
|
Print node and children, not node, children and siblings.
|
|
|
|
|
|
+/- 2. Thanks to Robin for catching this.
|
|
windows settings.
|
|
This uses the xaml printing methods. It is likely that I need to go do a DirectX approach since the xaml
printing method could run into memory problems as it seems the architecture requires all the pages to
be rendered prior to the print job even starting. Currently checking on this with a contact at MS.
I wanted to get this code in place now since it is working and it has the needed framework for much of
what we want. A change to DirectX requires a changes in the solution file that may be complicated and so it
makes sense to do a split here with respect to providing printing support.
|
|
Also fix issue that occurs in Windows 8.1 with an unhandled exception
if a mouse scroll occurs before a document is even opened.
Fix issue with keyboard based update of page slider.
|
|
Image masks don't have a colorspace; check before dereferencing.
|
|
|
|
|
|
|
|
|
|
|
|
Proper use of mutex lock in areas where we multiple threads can
not be accessing the document or page level objects at the same
time. With this fix, multiple threads can be rendering different
display lists at the same time.
Also fix for use of page selection scroll bar so that it does not
render the page until the user lifts off the bar. This is the
same as the Android app and avoids smaller devices getting overloaded
if you do rapid scrolling in a document.
|
|
Thanks to Micha for spotting this.
|
|
When putting store objects into the store, ensure that they do cannot
collide across documents.
|
|
If header field is too large, warn, but don't bale out.
|
|
|
|
Was failing to correctly initialise scissors when extending the stack.
This caused the fill text to disappear in fts_15_109.pdf
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Avoids a SEGV on writing.
|
|
When stroking and filling in a single operation, we are supposed
to form the complete stroke+fill image, then blend it back, rather
than filling and blending, then stroking and blending.
This only matters during transparency, or with non-normal blend
modes.
We fix MuPDF to push a knockout group when doing such operations.
|
|
While looking at fts_09_0921, I spotted that the alpha values on images
and fills aren't being sent. Fix that here, together with broken colors
being sent for masks.
fts_09_0921 still renders badly due to the lack of support for blend
modes.
|
|
The luminosity flag and background color are currently ignored.
The clip stack optionally held in the null device is updated here to
be a container stack, together with a flags word (currently just used
to indicate the type of the container at the current place in the
stack), and a user value (used by the SVG device to stash the id for
the mask it's generating).
|
|
Really just a structure indicator as SVG 1.1 doesn't support
blending etc.
|
|
fts_09_0919.pdf shows up some silly mistakes in the clip stack handling
and in the handling of 0 sized pixmaps. Simple fixes.
|
|
Attempt to improve the clarity of the explainations pertaining to
cloning contexts and bound contexts.
|
|
fz_reset_gel fails to reset the length of active edges, which could
(AFAICT) lead to pointers in gel->active pointing to memory that's
been previously freed by fz_resize_array.
|
|
A gradient fill that doesn't fill the bbox should be see through
(unless background color is set, but we'll worry about that case
when I find an example file that uses it). Arrange for the pixmap
we draw the gradient fill into to be transparent initially.
Also ensure that when we convert to png we preserve transparency.
|
|
|
|
For consistency with the rest of the code.
|
|
|
|
fz_clip_path takes a rect parameter, but all the callers of it use
NULL. In most cases they have a perfectly reasonable value that they
could pass to hand anyway. Update the code to pass this value, which
saves the need for the scissor stack keeping code to recalculate it.
|
