| Age | Commit message (Collapse) | Author |
|---|
|
https://codereview.chromium.org/2482663002/ )
Reason for revert:
Max cmsChannelsOf() is 15, which is larger than expectation of existing code and cause crashes (at least the fuzzer).
BUG=chromium:663240
Original issue's description:
> Clean up fx_codec_icc.cpp
>
> Committed: https://pdfium.googlesource.com/pdfium/+/a94fc11866adb1b9ca4a4e1afb4fb574ed472e07
TBR=dsinclair@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
Review-Url: https://codereview.chromium.org/2485363002
|
|
Review-Url: https://codereview.chromium.org/2482663002
|
|
Review-Url: https://codereview.chromium.org/2477443002
|
|
An optimization to speed up float-to-int rounding caused a different
result for one input value. This tweaks the conversion constant so
that the results are identical across the entire valid range, and
adds a test that checks the part of the range that is most sensitive
to errors.
BUG=pdfium:624
Review-Url: https://codereview.chromium.org/2466203002
|
|
BUG=pdfium:611
Review-Url: https://codereview.chromium.org/2381063002
|
|
TEST=build pdfium and chromium
BUG=pdfium:599
Review-Url: https://codereview.chromium.org/2355523002
|
|
FXSYS_round is painfully slow on Windows. It does range checking and
then calls an extremely expensive function. It ends up consuming half
the CPU time when decoding the images in PDFs such as this one:
https://www.ets.org/Media/Tests/GRE/pdf/gre_research_validity_data.pdf
SSE can be used to optimize this:
__m128 cmyk = {c * 255, m * 255, y * 255, k * 255};
uint32_t output[4];
_mm_storeu_si128((__m128i*)output, _mm_cvtps_epi32(cmyk));
but is cryptic, only works for x86/x64, and gives basically identical
performance to this solution - int(c * 255 + 0.5f);
The rounding behavior is not identical but in practice this rarely
matters, and in this specific case it does not matter because the edge
cases that vary are not hit.
The three divisions at the end were changed to multiplies because
profiling showed they were a significant cost.
This change reduces the image-decode stalls in the PDF listed above by
about 40%, making for a noticeably better experience. Further
optimizations are possible but would require significantly more time and
testing.
BUG=617365
Review-Url: https://codereview.chromium.org/2096723003
|
|
This change contains files in core directory which were not covered
in part I. This is part of the efforts to make PDFium code compilable
by Clang chromium style plugins.
The changes are mainly the following:
-- move inline constructor/destructor of complex class/struct out-of-line;
-- add constructor/destructor of complex class/struct if not
explicitly defined;
-- add explicit out-of-line copy constructor when needed;
-- move inline virtual functions out-of-line;
-- Properly mark virtual functions with 'override';
-- some minor cleanups;
BUG=pdfium:469
Review-Url: https://codereview.chromium.org/2060913003
|
|
Review-Url: https://codereview.chromium.org/2032613003
|
|
Add missing helper function to CFX_ByteTextBuf to avoid the
anti-pattern CFX_ByteString(sBuf.AsStringC()), using the name
"Make" to indicate there's an allocation going on in this case.
Change some method arguments to take pre-existing ByteStrings where
possible.
Review-Url: https://codereview.chromium.org/1977093002
|
|
All of the ICodec_* interfaces had a single implementation. This CL removes
the interfaces and uses the concrete classes in their place.
BUG=pdfium:468
Review URL: https://codereview.chromium.org/1876023003
|
|
Review URL: https://codereview.chromium.org/1864153002
|
|
Review URL: https://codereview.chromium.org/1832173003
|
|
This CL moves the core/src/ files up to core/ and fixes up the include guards,
includes and build files.
R=tsepez@chromium.org
Review URL: https://codereview.chromium.org/1800523005 .
|
