Fixes the first issue raised in #4937.

Regression from #4828.

This forces UTF-8 encoding by default on newer versions of Windows,
which fixes issues with accessing files containing characters
outside of the system's default locale/encoding.

* Fix font name for type3 extraction from pdfs

* free name on failure

* simplify free

Restore the files that were installed to fontforge-extras
Debian package prior to the FontForge 2020 March Release.

Co-authored-by: Anthony Fok <foka@debian.org>

#4979)

* FAQ: Why is Element --> Build disabled (or gray)?

* FAQ: When I use Element --> Overlap --> Remove Overlap the glymp gets inverted (or white) on the overlapping sections. Why?

* Inform the reader that both components must be available.

* Fix typo's

* Fix typo's

* Update doc/sphinx/faq.rst

Co-authored-by: Jeremy Tan <jtanx@users.noreply.github.com>

* Update doc/sphinx/faq.rst

Co-authored-by: Jeremy Tan <jtanx@users.noreply.github.com>

* Update doc/sphinx/faq.rst

Co-authored-by: Jeremy Tan <jtanx@users.noreply.github.com>

Co-authored-by: Jeremy Tan <jtanx@users.noreply.github.com>

… conversion (#4999)

The clut is only present when antialiasing is turned on

Fixes #4959

Add support for a "dimensions" flag in SplineChar.importOutlines in the Python bindings and in the import dialog box that sets the width of the glyph from an imported SVG. This works only for files with no units and absolute dimensions in the "width" and "height" attributes of the SVG.

GNU Mach microkernel header files (gnumach-dev 1.8+git20201129) added to
struct i386_xfp_save in /usr/include/i386-gnu/include/mach/i386/fp_reg.h
a new "extended" field which happens to collide with the macro definition
"#define extended double" in FontForge, leading to FTBFS on GNU Hurd.

This issue may be circumvented by temporarily undefining the "extended"
macro before loading `<gio/gio.h>` and friends, and redefining the
"extended" macro right after.

Special thanks to @henrich, @ucko and @kilobyte for reporting the issue
and offering suggestions on a fix.

See also https://bugs.debian.org/877795

Fixes #3119

Fixed sign error when importing the ascent and descent from an SVG font.

Detailed explanation here:
#5028

…up list ending with an empty node (#5025)

* Fix some calls to GDraw to only call draw functions from expose

* Fill progress in expose only

* Use expose calls to redraw updated glyphs in fontview

* Fix drawing of guide lines on expose calls

* combinations.c: Fix drawing on expose only

* charview.c: fix another draw outside of expose

* ggdkcdraw.c: enable warning and debug assert on auto-paint

* Upgrade GitHub Actions

* Upgrade GitHub Actions

* fix build

Co-authored-by: Jeremy Tan <jtanx@outlook.com>

Co-authored-by: Jeremy Tan <jtanx@outlook.com>

Fixes #5045

Fixes security issue in #5059.

* Fix typo in decodestr()

The typo resulted in the generation of complete gibberish when decrypting glyph charstrings.

* Fix decryption of binary eexec ciphertext

A font's binary eexec ciphertext is followed by 512 clear text '0' characters which may be interspersed with white space characters.  Typically, there is a newline character preceding the first '0'.  The program was treating that newline as a final character of the ciphertext rather than as part of the white space.

I understand that the API's there are not guaranteed stable, but this
would help a lot with `sfdnormalize` project, as it givess me a way to
actually test that my library `sfdutf7` returns correct results.

See also <https://twitter.com/fr_brennan/status/1555879287840034817/>.

cc: @alerque

* Prevent approximation from going back and forth

* Calculate slope at the end based on to->prev, not from->next

…penTy… (#4097)

* Add option to ignore non-existent glyphs in substitution rules of OpenType feature file

When reference to non-existent glyphs are present in an OpenType feature file, merging the feature fails.
Add an option to skip the erroring out in merging feature file. This helps with feature files
of Indic scripts with many hundreds of ligatures; some of the glyphs may not be present in all fonts,
yet the feature file can be reused.

TODO: handle the case when an entire lookup list is invalid.

* Merge feature file: skip error with empty lookup lists when invalid substitutions to be ignored.
The callchain is:
    1 fea_ApplyLookupListPST         featurefile.c
    2 fea_ApplyLookupList            featurefile.c
    3 fea_ApplyFeatureList           featurefile.c
    4 fea_ApplyFile                  featurefile.c
    5 SFApplyFeatureFile             featurefile.c
    6 SFApplyFeatureFilename         featurefile.c
    7 LoadKerningDataFromMetricsFile splinesaveafm.c

* Fix python method 'mergeFeature' optional parameter passing

* Add test for skipping invalid substitution rules when merging OpenType feature file

* Update test 1017 to check 'sub <glyphclass> by glyph' case

* Change the flag name to 'ignore_invalid_replacement'

#5107)

Co-Authored-By: Fredrick Brennan <copypaste@kittens.ph>

Link at https://fontforge.org/docs/scripting/scripting.html#scripting-example for testsuite scripts directs to 
https://github.com/fontforge/fontforge/tree/master/test
it should use
https://github.com/fontforge/fontforge/tree/master/tests

…alog

Fixes #5119

…xpose

Fixes #5090

Fixes #5162

Regression from #4889

2b15c0e#diff-d15a109191a6ac28861dafcc1e3ee5a96544c68911601ac42e6462cad71aa58eL2879