Thu 13 Apr 2017 08:50:51 AM UTC, comment #13:
OK I've gotten rid of the mark anchors to the preceding character, and instead placed all these diacritics to the top of the bounding box (and the low diacritics likewise to the bottom of the bounding box.)
The extreme example ʗ̃͡ʔ still crashes. The only conceivable solution would be a contextual positoning. I would not undertake that without fully understanding the scope of the project.
(We're talking about U+035C-62 -- and don't forget U+1DFC.)
|
Thu 13 Apr 2017 08:13:41 AM UTC, comment #12:
Hi David,
On further experimentation and reflection, I removed the 'mark' positioning on the double-width high diacritics, opting instead to just place them very high.
For the problem of placing a mark above two letters of different heights, the only mechanism OpenType has to offer is contextual positioning. But if you add to that further marks over the individual letters -- it has gotten out of hand, practically.
I made one concession to your case of ʗ̃͡ʔ by putting a 'mkmk' on the double-width mark, but as I was doing it, I thought this will cause a problem in other cases, and sure enough ... that causes a collision in ɪ̈͡β̞.
|
Wed 12 Apr 2017 10:51:48 PM UTC, comment #11:
OK I made some improvements, I hope, to Mono and Serif.
Some changes didn't take, somehow, in Mono.
Sans will have to wait for now.
|
Wed 12 Apr 2017 06:28:10 PM UTC, comment #10:
U+0361 COMBINING DOUBLE INVERTED BREVE is used in the IPA, the ALA-LC romanization of Cyrillic, and the transliteration of bind runes, among other things. There’s basically no limit to the possible combinations. Making it work with all combinations of IPA characters with IPA diacritics should suffice. Since that is a lot to ask, I got a few dozen examples of U+0361 from Wikipedia and checked them in FreeFont. Most sequences like fine, but in some of them the inverted breve overlaps one of the other characters.
FreeMono: ʈ͡ɽ̝̊
FreeMono Bold: ʈ͡ɽ̝̊
FreeMono Oblique: b͡β e̞͡β̞ i͡β̞ j͡β̞ k͡ʟ̝̊ ɘ͡β̞ ɛ̝͡β̞ ɨ̞͡β̞ ɪ̈͡β̞ ʈ͡ɽ̝̊
FreeSans: c͡ʎ̥
FreeSans Bold: c͡ʎ̥ ʈ͡ɽ̝̊
FreeSans Bold Oblique: c͡ʎ̥ ɪ̈͡β̞ ʈ͡ɽ̝̊ ʔ͡h ʕ͡ʡ ʡ͡ʜ ʡ͡ʢ
FreeSans Oblique: c͡ʎ̥ e̞͡β̞ m͡b q͡ǂ ɘ͡β̞ ɛ̝͡β̞ ɡ͡ǂ ɢ͡ǂ ɴ͡ǂ
FreeSerif: ʗ̃͡ʔ
FreeSerif Bold Italic: m͡b n͡d p̪͡f p͡f p͡ɸ q͡ǂ s͡ɸ v͡f ŋ͡ǂ ȹ͡f ɖ͡ɽ̝ ɖ͡ʐ ɘ͡β̞ ɛ̝͡β̞ ɟ͡ʝ ɡ͡ǂ ɢ͡ǂ ɨ̞͡β̞ ɬ͡s ɴ͡ǂ ʃ͡ɬ ʡ͡ʜ ʡ͡ʢ
FreeSerif Italic: m͡b p̪͡f p͡f p͡ɸ s͡ɸ v͡f ȹ͡f ɖ͡ɽ̝ ɖ͡ʐ ɛ̝͡β̞ ɢ͡ǂ ɬ͡s ɴ͡ǂ ʃ͡x ʃ͡ɬ ʈ͡ɽ̝̊
Some of these overlaps are more significant than others.
|
Sun 22 Feb 2015 09:26:28 PM UTC, comment #9:
Hi Hsing-yen He,
Let me repeat: my difficulty in fixing this is, I have no information about which characters are combinded with this tie.
Until I know that, I cannot do very much with this.
Thanks!
|
Sun 22 Feb 2015 04:40:40 PM UTC, comment #8:
I have not receive your e-mail from Gmail but found this release at launchpad.net, I think maybe your e-mail was broken by GFW, because CPC authority had broken Gmail ports last year.
|
Mon 16 Feb 2015 03:55:03 PM UTC, comment #7:
I switch to FreeSans, and I found this font still have problem for these combinations.
|
Mon 16 Feb 2015 03:49:03 PM UTC, comment #6:
Oh~sorry for my fault, I want to say "I am using 8 Feb. 2015 release".
|
Mon 16 Feb 2015 03:44:20 PM UTC, comment #5:
Hi Steve,
I copied some texts from here to Unicode Text Styler (http://babelstone.co.uk/Unicode/text.html) for test, they looks very well for me. I am using 8 Feb. 2014 release.
|
Sun 15 Feb 2015 01:15:52 PM UTC, comment #4:
Hi,
I had one other thought: maybe you haven't been getting my e-mails at all. You should have gotten (I think) two e-mails, not from the Savannah site but from my private e-mail account.
I also sent an e-mail with a link to the the Google Drive directory.
All these mails to the address listed for your Savannah account. Did you get them?
|
Sat 14 Feb 2015 11:43:56 AM UTC, comment #3:
Hi Hsing-yen,
As I said, I can see the problem.
I just have no way of knowing the extent of the use of this combining letter. Until I know that, I'm hesitant to start trying to "fix" it.
To exactly which combinations of letters is this tie applied?
|
Fri 13 Feb 2015 05:03:47 PM UTC, comment #2:
Hi Steve,
I have copied some IPA transcriptions from Wikipedia articles (starting from “International Phonetic Alphabet”) to test before I submit this item. I also found these articles referenced some professional phonetic documents such as the Handbook of the International Phonetic Association to explain.
|
Thu 12 Feb 2015 09:13:55 PM UTC, comment #1:
Hi Hsing-yen.
Well, I see what you mean of course.
Part of the problem here is that we have little in the way of examples.
If you could provide an image from (whatever kind of) a book which shows this combining character being used, that would be very helpful.
But this looks rather messy. A solution might take some time.
Another problem is... with which characters exactly does the tie interact?
|
Thu 12 Feb 2015 06:29:36 PM UTC, original submission:
When I use FreeSerif to display these combinations, The ligature tie (U+0361) will block some part of letter, or positioned aslantly:
p̪͡f, b̪͡v, t͡s, t͡θ, t͡ʃ, ʈ͡ʂ, ɖ͡ʐ, ɡ͡b, t͡ɬ, c͡ç, ɟ͡ʝ, c͡ʎ̝̥
All combinations with ǃ, ǀ, ǁ, ǂ, ʘ (such as ǃ͡q, ǀ͡ɢ, ǁ͡qʰ, k͡ǂ, ɡ͡ǂ, ǂ͡k, ʘ͡q)
|