| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Drop the unused 'serif' argument to the CJK lookup functions.
Use the BCP 47 names for CJK scripts and languages:
zh-Hant for traditional Chinese,
zh-Hans for simplified Chinese,
ja for Japanese,
ko for Korean.
The lookup function also allows commonly used language+country codes:
zh-TW and zh-HK for traditional Chinese,
zh-CN for simplified Chinese.
|
|
Using #ifdef FZ_ENABLE_ means we build code in, even if we have
defined FZ_ENABLE_WHATEVER to be 0 (as we do in config.h).
|
|
MuPDF may attempt to load a page but fail to do so, e.g. due to
a circular page tree. When this happens the page will never be
introduced into the document's list of pages. Its next and prev
pointers are both NULL, but the code in fz_drop_page() falsely
assumed that the prev pointer was always set.
Thanks to oss-fuzz for reporting.
|
|
|
|
Keep a list of currently open pages for each document. Attempting to
load a page that is already loaded will return the same instance again.
|
|
|
|
There is a regression for 2325_-_JPX_image_with_padding_rejected.pdf.
Object 3 in that document is a JPX-encoded image. Its EOC marker is
preceded by two extra bytes of data, 0x80 0x80. This makes the file
broken according to the JPEG 2000 specification.
Acrobat Reader and the Kakadu JPX decoder accepts this file without
issues, so OpenJPEG 2.1.0 added code to fix this (bug 226, commit
005e75bdc). That fix detects exactly two bytes of 0x80 0x80, a rather
brittle fix. Adding more padding or changing the padding byte values
is not accepted. Adding more padding is acceptable to Acrobat Reader
and Kakadu. An unrelated fix for another problem has since broken
OpenJPEG's support for this broken image.
|
|
This removes the need for unistd.h, which isn't around on
VS2005.
Also remove unused variable.
|
|
|
|
The upsampling code in the JPX decode attempted to guess a
suitable upsampling factor. The guessed factor was wrong,
causing writes of samples outside of the decoded image buffer.
Simply limiting the coordinates to the image buffer would
not suffice because the factor was wrong for every upsampled
row of pixels. openjpeg does provide an upsampling factor,
so use that instead and also take the component offsets into
account when decoding components into the pixmap. Combined
this resolves the issue that previously triggered ASAN.
Thanks to oss-fuzz for reporting.
|
|
pdf_page_transform() may throw due to a cycle in the page tree.
When this happened mupdf would previously forget to drop the
default colorspaces obtained, after this commit they are dropped.
Thanks to oss-fuzz for reporting.
|
|
|
|
|
|
|
|
Not popping causes assert to be triggered in fz_draw_end_group().
|
|
|
|
|
|
Not popping causes assert to be triggered in fz_draw_end_group().
|
|
|
|
|
|
This makes it possible to change the colorspace when encountering
ICC colorspaces.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Unlikely that this could be an issue, but lets add it for safety.
|
|
|
|
Previously there was no visibility as to what the error was.
|
|
fz_fill_path() may throw an exception halfway through
pdf_show_path(), which in this case would not attempt to end any
begun groups or softmasks. This led to e.g. leaks of pixmaps held
by a group that was never ended.
Moving the cleanup to the always block is not foolproof because
the cleanup code itself may also throw exceptions, hence
preventing the end of the fz_always block from being executed.
This commit does put pdf_show_path() in the same situation as
pdf_run_xobject() that has the same problem with its cleanup
code.
Thanks to oss-fuzz for reporting.
|
|
fz_open_jbig2d() is called at two locations in MuPDF. At one
location a reference to the JBIG2 globals struct was taken before
passing it to fz_open_jbig2d(). At the other location no such
reference was taken, but rather ownership of the struct was
implicitly transferred to fz_open_jbig2d(). This inconsistency
led to a leak of the globals struct at the first location.
Now, passing a JBIG2 globals struct to fz_open_jbig2d() never
implictly takes ownership. Instead the JBIG2 stream will take a
reference if it needs it and drops it in case of error. As usual
it is the callers responsibility to drop the reference to the
globals struct it owns.
|
|
JBIG2 images are detected by build_compression_params() and then
always passed to fz_open_image_decomp_stream() by build_filter().
Therefore there is no chance for build_filter() at a later stage
to detect JBIG2 images, and so that check can be removed.
|
|
|
|
When painting images using interpolation 16.16 fixpoint
arithmetics is used. This limits the width/height of any
image that can be painted to 32767. There was no size check,
so large images caused overflow and subsequent out of bounds
accesses which triggered MSAN. This c
Thanks to oss-fuzz for reporting.
|
|
Previously this would result in trying to dereference a NULL pointer.
Thanks to oss-fuzz for reporting.
|
|
Previously a value of 0 would cause the document not to be rendered.
|
|
The signer object now has an extra method that informs the caller of
the maximum size the digest might be. This is used to allocate space for
the digest within the file and to size some of the buffers used in the
code.
The openssl-based inplementation of the signer object has been updated to
perform a test digest generation so as to find the size needed. We
believe that the digest size is indendent of the hashed data.
|
|
|
|
Avoid pdf_pass_event interacting with widgets that have the read-only
field set.
|
|
pdf_show_pattern() may resize the pdf run processors gstate array. In doing
so any pointers to it saved by called will become stale. Similarly to
pdf_begin_group() pdf_show_pattern() now returns the updated pointer to the
top of the gstate stack.
For the same reason pdf_show_pattern() can not be given a pointer to a gstate
stack entry and continue using it after the gstate array has been resized. So,
have pdf_show_pattern() take an index to the desired state instead.
|
|
The leak was triggered by the PDF from 699576, but this
commit does not fully fix that bug.
|
|
This was used to prevent infinite cycles in the Type3 font CharProcs,
but the previous commits forbids all cycles in order to prevent memory
leaks due to cyclic reference counts.
|
|
Thanks to oss-fuzz for reporting.
|
|
build_filter_chain_drop() promises to extend (according to the
fs argument) the filter chain it is given, or in case of exception
throw away the at that point potentially extended filter chain it
was given from the beginning.
Because build_filter_chain_drop() calls build_filter_drop() for
every filter it adds it doesn't need to do any cleanup of the
filter chain on its own, that's build_filter_drop()'s responsibility.
Prior to this commit fz_catch() in build_filter_chain_drop() which
would drop the filter chain one time too many (it was already dropped
by build_filter_drop()), causing the callers to use a stale pointer.
Now once the extra fz_drop_stream() has been removed the logic works
as it ought to, even in the case of exceptions. Thanks to oss-fuzz
for reporting.
|
|
Thanks to oss-fuzz for reporting.
|
