The character is clipped if it is decorated by a combining character. Here is an example with IPA diacritics in Firefox (font: Segoe UI, backend: cairo):

Tried ClipBoxFix = 1, it doesn't work. I think this is caused by the combining character. Since its clipbox is smaller than the decorated character's, part of the decorated character is clipped.
try set clipboxfix in experimental section not as per-app basis. And not the last section in the profile. There are some known bugs related to ini parser.
tried, nothing happened _(:蟹銆嶁垹)_
I think this is a similar issue. This is with ClipBoxFix=1

@snowie2000 Anyway to resolve the above issue when combining characters?
Please provide me a piece of text to test.
I guess I'll need to give you a piece of text and font, as this is only an issue when font ligatures are enabled.
@snowie2000 The font is PragmataPro. Is there a way for me to upload it to you privately? Since this is not a free font I do not want to upload it here.
You can send it to my email. Check my github profile.
@snowie2000 done!
@tcbbd hey, is this still an issue with the latest MacType?
@snowie2000 you've mentioned no further changes to clipboxfix ... want me to close or leave open?
@sammilucia I almost forgot it :upside_down_face:
Alright, I walked through my mailbox and found the mail about the issue, I mean, the issue about the incorrect rendering of "font ligatures" instead of the real topic "diacritics".
Actually, I've never heard back from @tcbbd, so there is not much I can do.
As of the ligature problem, I don't see any defect on my webstorm with the latest mactype. No ghost shadow, no missing letters.
@IceMan81 Did I do anything wrong in my configuration?

I don't see the problem on my PC either with a quick look around (mine used to be mainly on Control Panel stuff)




Going to close this now so we can keep up momentum, but please let us know if you're still having it with beta2 @tcbbd @IceMan81 馃槉
@snowie2000 thanks for trying out the font. Your test is correct.
Looking at your screenshot I see the shadows in the double equals, triple equals and the greater than symbols. See how they look when mactype is turned off. You should see a difference.
@sammilucia do you mind sharing your ini file. Your rendering looks really good. Not sure if it鈥檚 because I鈥檓 viweing your screenshot on a phone!
haha of course! I've actually snuck it into the next beta3 ;)
here 'tis 馃槉
@snowie2000 thanks for trying out the font. Your test is correct.
Looking at your screenshot I see the shadows in the double equals, triple equals and the greater than symbols. See how they look when mactype is turned off. You should see a difference.
Hmm, that is not a shadow, it's the result of vertical antialias.
This is what it looks like under cleartype:

Pretty clean, right?
Let's see what does the font look like in real

And this is what it looks like in large size:

I believe you already got it.
I see, I think I get what you are saying, but why would I not see this on a Mac then?
I have no idea how mac os does with it, maybe the vertical AA is off for small size fonts?
Well, what I'm talking about is the Clipbox problem. The clipbox is sometimes smaller than the glyph, causing the imperfect rendering results. It can happen in many circumstances, e.g., for combined letters (maybe because the clipbox of the combined letter is miscalculated using metrics of the diacritic symbol), or for nastaliq arabic letters (see https://ur.wikipedia.org). IIRC this can also happen for chinese fonts (or for latin letters within chinese context, I'm not sure), but I don't have an example at hand.
It seems you're trying to fix the problem with clipbox, so I assume this issue has already been confirmed.
Here's what i see on the Mac.

You're using such a large sized font, you won't see any shadow in Windows too.

The problem is a lot less in some size due to the pixel alignment.

Well, what I'm talking about is the Clipbox problem. The clipbox is sometimes smaller than the glyph, causing the imperfect rendering results. It can happen in many circumstances, e.g., for combined letters (maybe because the clipbox of the combined letter is miscalculated using metrics of the diacritic symbol), or for nastaliq arabic letters (see https://ur.wikipedia.org). IIRC this can also happen for chinese fonts (or for latin letters within chinese context, I'm not sure), but I don't have an example at hand.
It seems you're trying to fix the problem with clipbox, so I assume this issue has already been confirmed.
@tcbbd I need precisely the letters and the program that will cause this issue to narrow down the problem 'cause I know nothing about the diacritic symbol and I don't know how to type them out.
@snowie2000 OK, I can see the difference between the different font sizes, same as what you see. Size 28 onwards its very clear.
@snowie2000 Here is an example:
a蜏b蜏c
There is a "combining double breve below (U+035C)" between a and b, and another one between b and c. They are shown in Chrome, but not in Firefox (with DirectWrite disabled and MacType). This is an extreme example in which the diacritic symbol is completely disappeared since it is below the base line, most cases are not like this and maybe not so noticeable, but this definitely indicates some problem in Mactype.
Confirmed, it did happen.
Double confirm with my font tester demo.
Worse than I thought.

Hmm, you may have guessed it wrong, the one on the left is MacType. (Does it look more reasonable? :laughing: )
@IceMan81 i've just changed the italics slant to -2, it was too strong on fonts without their own native oblique/italics files.
@sammilucia Thanks for the update. Actually, I never understood what the ItalicSlant was meant to do to a slanting typeface.
@IceMan81 so now you know! 馃槀
After a difficult journey investigating the logic behind the GDI diacritics handling, I finally figured out that this feature requires OpenType layout support which is, sadly, removed from FreeType2.
FreeType suggests using a third party library to do the job. So I decided to put this feature/bug down for a while and leave it to the next stage.
The next release will not include a fix for this diacritics problem unless a corresponding fix is contributed. Further investigation is planned.
Good decision, I think
Most helpful comment
Good decision, I think