Glyph insertion is an expensive operation and we like to have it limited
based on buffer's input size which is handled by buffer's max_ops.
clusterfuzz-testcase-minimized-harfbuzz_fuzzer-5754958982021120:
Before the change: 0.67s user 0.00s system 99% cpu 0.674 total
After the change: 0.02s user 0.00s system 98% cpu 0.024 total
Which takes much longer on valgrind and tsan bots.
../test/fuzzing/hb-set-fuzzer.cc: In function ‘int LLVMFuzzerTestOneInput(const uint8_t*, size_t)’:
../test/fuzzing/hb-set-fuzzer.cc:38:82: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]
38 | const instructions_t &instructions = reinterpret_cast<const instructions_t &> (data);
|
And it is already disabled at project level so let's disable it here also.
Instead of passing dependencies as required we used one giant shared
dependency list containing all dependencies for every library/executable.
While this kinda works, the specified deps are also used for generating
the pkg-config files and this leads to lots of Requires.private and Libs.private
entries which aren't really needed.
This removes the "deps" array and replaces it with a few smaller ones and
makes sure the public libraries only get passed the dependencies actually
needed.
Fixes#2441
So one can run a category of interested tests like
meson test -Cbuild --suite aots --suite src --print-errorlogs
Intead issuing particular tests which also is possible like
meson test -Cbuild test-shape --print-errorlogs
AAT::Lookup has no other way to detect whether it is returned from
a real and sanitized font data or from a null pool, this checks if
the table has been recognized valid by sanitizer by checking
table's major version which is zero if returned from a null pool and
non-zero if is from a sanitized font data, it is expected the other
calls of the table (unlikely to have more calls however) also do a
similar version check before calling the lookups used on the table.
Fixes https://crbug.com/oss-fuzz/21560
revert () does not clean up useless object_t. Adjust the order of
subsetting substitutions and conditions to avoid dangling object_t.
'i' shouldn't become equal to array's length which as the increament
is happened at end of the loop, if the input is bigger than all the
table coords, it will be equal to array's length.
Fixes https://crbug.com/oss-fuzz/21092
Use add_range instead an inner loop, clamp its input number by
number of glyphs a face has.
Even the face cmap12 and 13 have 32-bit hb_codepoint_t, which is here
used to make timeout, face's maxp has 16-bit gid limitation at least for now,
using that makes sure we both fix and the timeout and don't need to change
much things here also in order to support 32-bit gids also someday.
Fixes#2204
This specially becomes concerning on sub-components where a gvar table
that is sanitized using maxp's glyphs number overflows when a high gid
accepted here goes to it, maybe an additional check can be put there
also, this however feels to be enough.
Fixes https://crbug.com/oss-fuzz/20944
They aren't contributing to rendering and making issue for stroking, let's skip them
ourselves as Skia does also https://skia-review.googlesource.com/c/skia/+/268166
They are useful for extracting extents and so which that functionality won't be effected by this change.
Specially checks correctness of the API semantics:
* no move happens when a path is already opened with move-to.
* no path will be left open and close-path will happen at the end of opened paths.
* no path opens with a move-to and will be closed with no length.
* paths start and ending points matches.
* no line/quadratic/cubic command will be issued when no path is started.
Newer versions of MSVC with /we4146 don't like putting negative sign behind a
unsigned number as https://github.com/harfbuzz/harfbuzz/pull/2069
That however have made https://crbug.com/1050424 this complain:
src/hb-ot-color-sbix-table.hh:304:28: runtime error: negation of -2147483648 cannot be represented in type 'int';
cast to an unsigned type to negate this value to itself
which apparently can be fixed using this change.
Let's see if this won't make another ubsan complain!
fixes https://crbug.com/oss-fuzz/19956
am not super happy with the fix, guess we should do some check
before the memcpy anyway as @blueshade7 thinks also,
so let's have it or revert it when we have a better approach for the case.
Ebrahim Byagowi
Christoph Reiter
ckitagawa
ariza
Qunxin Liu
Garret Rieger
Khaled Hosny
Tim-Philipp Müller
Mathieu Duponchelle
ckitagawa-work