In ImageOutputDev it comes directly from malloc, and malloc guarantees
alignment for basic types, so we're good

In ArthurOutputDev it comes from QImage::bits that uses malloc
internally, so we're good

In cairo* it comes from cairo_image_surface_get_data that comes from
pixman_image_get_data that returns a uint32_t * so we're only going to
the original type alignment

oss-fuzz/10284

oss-fuzz/12638

oss-fuzz/8612

Makes for easier understanding/scoping of its value

oss-fuzz/8715

https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=8516

https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=8798

Sounds rather scary since we're storing a bigger value than what really
fits.

Fixed by the suggestion of https://developer.gnome.org/glib/stable/glib-Date-and-Time-Functions.html#GTime

Changing the type of _PopplerAttachment ctime/mtime would change the
structure size, thus break the BC, so leaving that for the future for
now

oss-fuzz/12608

https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=8438

Contains code inspired in code by Hans-Ulrich Jüttner and Adrian Johnson

poppler/poppler#709

This is nowadays also provided by CMake itself via CMakePushCheckState.

Creates a map of all the named destinations in the document.

The information about a destination used in a PDF document.

The availability depends on the _FILE_OFFSET_BITS define, which is set
in config.h. So we need to have this during the configure check as well,
to test under the same conditions.

Add a method the check if an outline item has children to avoid having to eagerly expand the hierarchy anyway.

Port the Qt5 viewer demo to use the lazy outline item API (even if for loading all items up front for now).

Remove the intermediate Outline type since all items are owned by the document and the top-level items will always be eagerly loaded anyway.