There were two issues:
1) the C+=R and C-=R operators expected arithmetic types to have .real()
2) the R+C, R-C, R*C and R/C operators applied the operation in wrong
order (did C+R, C-R, C*R and C/R instead). This wouldn't matter for
+ and * which are commutative, but is incorrect for - and /.
We were calling write syscall twice for every sample, which effectively
hurt the writer's performance.
With this change exporting a melody in the Piano app now takes less than
a second, which previously took about 20 seconds on my machine.
Additionally, I've removed an unused `WavWriter::file()` getter.
This will make window.open a lot easier to implement. As written, the
implementation of Navigable::choose_a_navigable now looks a lot closer
to the old BrowsingContext::choose_a_browsing_context. With the notable
exception that we still crash in many cases, and don't properly handle
multiple top-level traversables in the same WebContent process.
We currently do not wait for iframes to finish loading before triggering
the document's load event, which creates a race condition for any ref
tests that include iframes. Until that gets fixed, let's skip the one
affected test.
See issue #22012.
Ideally we would not create a layout node at all for these elements so
that every layout node would always have a paintable associated with it.
But for now, to fix the crash, just leave a FIXME and special case this
element.
Also leave a VERIFY to make it easier to debug this type of crash in the
future.
Fixes a crash seen on codecov.io for my 'patch' project.
SMasks are greyscale images that get used as alpha channel for a
different image.
JPEGs in PDFs are stored as streams with /DCTDecode filters, and
we have a separate code path for loading those in the PDF renderer.
That code path just calls our JPEG decoder, which creates bitmaps
with format BGRx8888.
So when we process an SMask for such a bitmap, we have to change
the bitmap's format to BGRA8888 in addition to setting alpha values
on all pixels.
It's perfectly possible for JavaScript to call unobserve() on an element
that hasn't been observed. Let's stop asserting if that happens. :^)
Fixes#22020
Before this change, there was some confusion possible where an IO would
try to find its way back to the document where we registered it.
This led to an assertion failure in the test I'm adding in the next
commit, so let's fix this first.
IOs now (weakly) remember the document where they are registered, and
only unregister from there.
Per 1.7 spec 3.8.1, there are multiple logical text string types:
* text strings
* ASCII strings
* byte strings
Text strings can be in UTF-16BE, PDFDocEncoding, or (since PDF 2.0)
UTF-8.
But byte strings shouldn't be converted but treated as binary
data.
This makes us no longer convert strings used for drawing page text.
TABLE 5.6 "Text-showing operators" lists the operands for text-showing
operators as just "string", not "text string" (even though these strings
confusingly are called "text strings" in the body text), so not doing
this there is correct (and matches other viewers).
We also no longer incorrectly convert strings used for cypto data
(such as passwords), if they start with an UTF-16BE or UTF-8 marker.
No behavior change for outlines and info dict entries.
https://pdfa.org/understanding-utf-8-in-pdf-2-0/ has a good overview of
this.
(ASCII strings only contain ASCII characters and behave the same
anyways.)
Manually added an Outlines dict with three items, one each for
every text string encoding in its title.
(Preview.app apparently can't handle UTF-8 in outlines either.)
For now, this uses UTF-16BE and UTF-8 marked strings in page body text.
These markings should be ignored in body text.
Hand-written, with `set fenc=latin1` and `set binary` in vim, and
xref etc fixed up by running
mutool clean Tests/LibPDF/encoding.pdf Tests/LibPDF/encoding.pdf
as usual.
Else, outline items that have newlines in them end up with a weird
vertical offset.
(This does affect the outline item's tooltip, which shows the whole
title. But not having a newline there seems alright, arguably
preferable.)
The title of an OutlineItem is already in UTF-8.
This is currently done in LibPDF's Parser::parse_string(). I think
that's not quite the right place (it shouldn't be done for all strings)
and not done quite right (text strings should convert from
PDFDocEncoding to UTF-8 unless prefixed by an UTF-8 BOM), but even if
that changes, I think we'll keep OutlineItem.title in UTF-8.
Implemented by adding the extra 3-value syntax as its own case and only
running it when parsing background-position. I'm sure it could be
implemented in a smarter way but this is still a bunch less code than
before. :^)
This means `object-position` will no longer incorrectly accept the
3-value background-position syntax.
Remove the now-ambiguous and unused `position` enum while we're at it.
(This enum only existed as a hack.)
