Added auto hinting compensation to fix too wide gaps between small glyphs

[oomek]

• Has this change been discussed on the forum or in an issue before?
• Does the code follow the SFML Code Style Guide?
• Have you provided some example/test code for your changes?
• If you have additional steps which need to be performed list them as tasks!

---

Description

Forced Autohint requires left and right bearings to be taken into acccount during kerning calculation. Without this the gaps between small glyph sizes are twice as wide on a lot of fonts.

This PR is related to the issue #1642 (comment)

Tasks

• Tested on Linux
• Tested on Windows
• Tested on macOS
• Tested on iOS
• Tested on Android

[LaurentGomila]

Looks ok, except some minor adjustments

[oomek]

Regarding that added function parameter:

SFML/include/SFML/Graphics/Font.hpp

Line 220 in d2406fc

float getKerning(Uint32 first, Uint32 second, unsigned int characterSize, bool bold = false) const;

Shall I update the brief as well? If so, do we need to indicate that when different bold flag is used than already cached the glyph will be loaded?

[eXpl0it3r]

Shall I update the brief as well?

If you think the brief could need an update/change, then certainly.
What you certainly need to add is /param for the new parameter.

If so, do we need to indicate that when different bold flag is used than already cached the glyph will be loaded?

Does the user of the function need to care about this implementation detail?
If only in some very special cases, then no, I wouldn't add that information.

[oomek]

Maybe it's just me being paranoid, but I found 2 fonts out of 50 I tested with unusually narrow whitespace.
It's not relevant to kerning compensation, but how whitespaceWidth is calculated. Maybe it's also a good opportunity to address it aswell by clamping the whitespaceWidth to at least 2 pixels by:

    float whitespaceWidth = m_font->getGlyph(L' ', m_characterSize, isBold).advance;
    if (whitespaceWidth < 2.f) whitespaceWidth = 2.f;
    float letterSpacing   = ( whitespaceWidth / 3.f ) * ( m_letterSpacingFactor - 1.f );
    whitespaceWidth      += letterSpacing;

Here is the screenshot of the affected font before and after:

Please open it in 100% browser zoom to avoid any potential scaling.

Here is the current state for comparison:

[oomek]

Would you like me to commit that whiteSpaceWidth compensation as well?

[LaurentGomila]

Since it's a different issue, let's address it in another PR.

[oomek]

Half of the CI tasks did not start. Do they have to be triggered manually?

[eXpl0it3r]

Half of the CI tasks did not start. Do they have to be triggered manually?

Our own CI setup only executes on demand on third-party PRs for security reasons.
I have since triggered them.

@LaurentGomila any additional comment for the PR?