Pillow/docs/releasenotes/4.2.0.rst
Jon Dufresne eb460a2ac8 Fix internal Sphinx link to installation doc
Fixes linkcheck warnings:

    handbook/image-file-formats.rst:563: [broken] ../installation.html:
    releasenotes/4.2.0.rst:7: [broken] ../installation.html:
2020-10-17 11:23:58 -07:00

52 lines
1.7 KiB
ReStructuredText

4.2.0
-----
Added Complex Text Rendering
============================
Pillow now supports complex text rendering for scripts requiring glyph
composition and bidirectional flow. This optional feature adds three
dependencies: harfbuzz, fribidi, and raqm. See the :doc:`install documentation
<../installation>` for further details. This feature is tested and works on
Unix and Mac, but has not yet been built on Windows platforms.
New Optional Parameters
=======================
* :py:meth:`PIL.ImageDraw.floodfill` has a new optional parameter:
threshold. This specifies a tolerance for the color to replace with
the flood fill.
* The TIFF and PDF image writers now support the ``append_images``
optional parameter for specifying additional images to create
multipage outputs.
New DecompressionBomb Warning
=============================
:py:meth:`PIL.Image.Image.crop` now may raise a DecompressionBomb
warning if the crop region enlarges the image over the threshold
specified by :py:data:`PIL.Image.MAX_IMAGE_PIXELS`.
Removed Deprecated Items
========================
Several deprecated items have been removed.
* The methods ``PIL.ImageWin.Dib.fromstring``,
``PIL.ImageWin.Dib.tostring`` and
``PIL.TiffImagePlugin.ImageFileDirectory_v2.as_dict`` have
been removed.
* Before Pillow 4.2.0, attempting to save an RGBA image as JPEG would
discard the alpha channel. From Pillow 3.4.0, a deprecation warning
was shown. From Pillow 4.2.0, the deprecation warning is removed and
an :py:exc:`IOError` is raised.
Removed Core Image Function
===========================
The unused function ``Image.core.new_array`` was removed. This is an
internal function that should not have been used by user code, but it
was accessible from the python layer.