| Age | Commit message (Collapse) | Author |
|---|
|
When repairing, if we hit a problem after having found a root object
live with that root object rather than giving up completely.
Also fix a memory leak, and cope better with trailing crap.
Thanks to Zeniko for these.
|
|
Thanks to Zeniko for pointing this out.
|
|
Thanks to Zeniko for pointing this out.
|
|
We can't do them properly at the moment, so don't try to do them at
all - only makes stuff worse for now.
Thanks to Zeniko for this.
|
|
Exactly parallel to the png problem fixed a few days ago. Thanks to
Zeniko for pointing this out.
|
|
In error cases, ensure we free objects correctly. Thanks to Zeniko
for finding the problems (and many of the solutions!)
|
|
In various places in the code, we add markers (".seen") to
dictionaries as we traverse them to ensure that we don't
go into infinite loops.
Adding a dictionary entry is bad as it's a) an expensive
operation, b) a potentially destructive one, and c) produces another
possible point of failure (as mallocs can fail).
Instead, add a flag to each dict to allow them to be marked/unmarked
and use that instead.
Thanks to Zeniko for pointing out various places that could usefully
be protected against infinite recursion.
|
|
Externalise free_link_dest, use it in doc_outline.c to avoid a memory
leak.
Thanks to again to Zeniko for pointing this out.
|
|
Only create a shape if we need to.
Correctly cleanup, thus avoiding a double free of mask later on.
Thanks to Zeniko for spotting these.
|
|
In pdfshow, if we fail to parse an object, just skip it rather than
aborting. Thanks to Zeniko for the suggestion.
|
|
The logic controlling whether to expand a trailer dictionary
or not was reversed. Fixed here. Thanks to Zeniko for pointing
this out.
|
|
|
|
|
|
A new 'cookie' parameter is added to page rendering/interpretation
functions. Supply this as NULL to get existing behaviour.
If you supply a non-NULL cookie, then this is taken as a pointer to
a struct that can be used for simple, non-thread locked communication
between caller and library.
The entire struct should be memset to zero before entry, except for
specific flags (thus coping with future extensions to this struct).
The abort flag should be zero on entry. It will be checked periodically
by the library - if the caller sets it non-zero (via another thread)
then the current operation will be aborted. No guarantees are given as
to how often this will be checked, or how fast it will be responded to.
The progress_max field will be set to an integer (-1 for unknown)
representing the number of 'things' to do. The progress field will
count up from 0 to this number as time goes by. No guarantees are
made as to the accuracy of this information, but it should be
useful for offering some sort of progress bar etc. Note that
progress_max may increase during the job.
In general, callers should be careful to accept out of range or
invalid data in this structure as this is deliberately
accessed 'unlocked'.
|
|
Add simple combined exe build for mupdf/muxps tools.
|
|
As we create clips/transparency groups etc, we maintain a stack of
draw device states.
Here we change this from having a 'current' state, and storing changes
on the stack to always keeping the complete current state on the head
of the stack. This should make error cleanup easier, as well as being
conceptually easier to follow (for me at least).
|
|
When starting to tile, create a shape plane if one exists.
When finishing tiling, plot the shape plane back too.
This solves the SEGVs. Something isn't quite right with the blending
colours on part of this file though.
|
|
Fix a 'TODO'. When asking for the bounds of a shading, simply
convert the shadings bbox and intersect with that.
|
|
In order for hyperlinks to work, we need to load the outlines after
the pages tree.
|
|
Testsseem to indicate that this works, and gives noticable
improvements. Enabling by default.
|
|
The xps png reading code was overwriting its context pointer with
zero, then referencing through it.
|
|
When the page number is given by an int, rather than an indirect
reference number in a link destination, the page number is already
0 based. Remove the adjustment that I'd been incorrectly doing.
|
|
When reading an XYZ link destination, I was writing Z over X. While
fixing this, change to cryptic enums rather than unexplained magic
numbers.
When reading the outlines out of a pdf file, better to actually store
them in a linked list rather than just dropping them.
|
|
The shape plane was being incorrectly offset from the dest plane.
This could have caused SEGVs or incorrect results when blending
back.
This solves the hivemind.pdf iphone being clipped on the right hand
side.
|
|
More work on MuPDFs blending support. This seems to get l16.pdf
working pretty well. Fuller testing required to know if we get
everything right, but the maths now makes as much sense as it
ever has to me, and seems to track what gs does (more or less,
in that we don't use group alpha and gs does).
|
|
|
|
Move 'kind' into the fz_link_dest structure (as this makes more sense).
Put an fz_link_dest rather than just a page number into the outlines
structure.
Correct parsing of actions and dests from pdf outlines.
|
|
As Zeniko correctly points out in bug 692593, I had incorrectly
used hp[n1] when I should be using hp[0] at various points.
Hopefully that bug should be solved in it's entirety now with
the previous commit.
|
|
Restrict images to the size of the shape when blending back.
This seems to solve the problem; leaving code disabled until full
tests complete.
|
|
|
|
|
|
Move to a non-pdf specific type for links. PDF specific parsing is
done in pdf_annots.c as before, but the essential type (and handling
functions for that type) are in a new file fitz/base_link.c.
The new type is more expressive than before; specifically all the
possible PDF modes are expressable in it. Hopefully this should
allow XPS links to be represented too.
|
|
gcc 4.4.5 gives helpful warnings about variables that can become
unset due to setjmp/longjmp usage. Fix that here.
Thanks to Sebras.
|
|
A couple of bits of the code SEGV in the event that values are NULL.
Fixed here by converting a do...while to a while, and adding an extra
guard in the if.
Thanks to Gaetan Bisson for the report, and patch.
|
|
|
|
|
|
Hadn't been updated to allow for context/exceptions.
|
|
One of the previous memsqueezing fixes (specifically that in
close_dctd) appears to cause the Memento fork bases squeezing
process to stop.
This appears to be because old code would do a NULL dereference
causing a SEGV. This would somehow NOT be picked up by the signal
handler, and the child would exit.
If the code is fixed to avoid the SEGV the code then somehow
goes on to do something (not in the close_dctd code) that makes
the mem squeeze process grind to a halt - but NOT in the same
instance of the executable. I am at a loss to explain this, but
would rather the code stays as it is (being as far as I can see
correct) for now.
|
|
|
|
Disabled for now by default.
Better SEGV handling under windows (to facilitate scripted squeezing)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Rather than passing a stream to a close function, just pass context
and state - that's all that is required. This enables us to
call close to cleanup neatly if the stream fails to allocate.
|
|
|
|
Mupdf has some large structures (fz_shade is over 32K!), so to catch
all pointers we need to search further than the 1K I was originally
limiting ourselves to.
|
|
|
