bugFree UCS Outline Fonts - Bugs: bug #45760, Положение...


bug #45760: Положение (отображение) диакритик 0300-0315 FreeSerif.

Submitted by:  Alexey M-ov <teorks>
Submitted on:  Fri 14 Aug 2015 02:31:01 PM UTC  
Category:  character range Severity:  3 - Normal
Item Group:  letter positioning poor Status:  Proceeding
Privacy:  Public Assigned to:  Steve White <Stevan_White>
Open/Closed:  Open Release:  development

Add a New Comment(Rich Markup)

You are not logged in

Please log in, so followups can be emailed to you.


( Jump to the original submission)

Mon 19 Dec 2016 03:48:26 PM UTC, comment #27:

Adding Acute over Cyrillic Ї as file #34651 mentioned is also required by Wikipedia article “Verkhovna Rada” (Ukraine’s Supreme Council).

Hsing-yen He <hammerandsickle>
Tue 28 Jun 2016 07:34:18 AM UTC, comment #26:

Before UTN #41 I have not found such a good guide for how to use them for Church Slavonic, they should be available without specific typeface to avoid some effects as bad mark posisioning or Tofu.

Hsing-yen He <hammerandsickle>
Sat 09 Apr 2016 04:12:11 AM UTC, comment #25:

The Unicode Technical Note no. 41 has a special section to explain the usage of diacritical marks in Church Slavonic. You can discovery some special positionings that have not implemented in FreeSerif.


Hsing-yen He <hammerandsickle>
Mon 07 Mar 2016 03:58:41 PM UTC, comment #24:

In 28-2-2016 build ế and ề works proper in non Vietnamese specific texts now, tested on both Firefox and BabelPad.

Hsing-yen He <hammerandsickle>
Fri 12 Feb 2016 03:11:20 PM UTC, comment #23:

Hello Steve,

Ukrainian I sometimes has dotted form (İ) in uppercase.


Is it possible of implementing it in FreeSans/FreeSerif ?

Hsing-yen He <hammerandsickle>
Fri 18 Dec 2015 07:45:37 PM UTC, comment #22:

Hello Steve,

I have found a online resource hosts some Old Cyrillic materials. You can see some special positionings as Aleksey refered if you want to make more improvements on Cyrillic script.


Hsing-yen He <hammerandsickle>
Mon 23 Nov 2015 12:14:43 PM UTC, comment #21:

Please, Hsing-yen. Your post again has nothing to do with the topic of this discussion.

Furthermore, this isn't a bug -- you are proposing something.
It might be best to discuss such things directly with me by e-mail.
(I looked at the file, but I can't really figure out what you're trying to accomplish.)

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Mon 23 Nov 2015 11:49:11 AM UTC, comment #20:

I made a proposal for a new style set that can generating Pinyin styled letters, all of them can be found in many textbooks and dictionaries published from PR China.


Hsing-yen He <hammerandsickle>
Sat 21 Nov 2015 10:08:19 PM UTC, comment #19:

Hello Alexey,

You have given us very little information to work with.

Images are worth a thousand words.

You built it with FontForge... do you mean you used the Makefile included in the project? If not, you have probably built it incorrectly.

Which version of the SFD files are you using? Which version of FontForge?

You say you are using it with Word '03. I assume this is on some version of Windows. Which version of Windows?

In Windows 7, it often happens that a font is incompletely deleted. If you open a DOS console, cd to C:\Windows\Fonts and type
dir Free*
you will see any stray copies. What do you see?

With the above information, we might have a chance at answering your question.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 21 Nov 2015 11:33:27 AM UTC, comment #18:

Есть такая проблема: когда я заново создаю шрифт из открытого в FForge, например, последнего FreeSerif (некорректно выглядит в Word), в новом шрифте символы в целом начинают корректно отображаться в Word'03, но диакритики в таком файле отваливаются и выглядят как ненастроенные (в отличие от оригинала), и мне приходится их настраивать заново своими методами, как я умею... Why!8-(

Alexey M-ov <teorks>
Sun 13 Sep 2015 01:08:47 PM UTC, comment #17:

Hi Steve,

I have read your posts again and I re-thoughted my last test. As my imagination, these Vietnamese forms in Latin script should performanced as same as Serbian forms in Cyrillic.

Hsing-yen He <hammerandsickle>
Fri 11 Sep 2015 06:17:14 PM UTC, comment #16:

Addtionaly, this app also have no way of indicating languge to text, much less corresponding font feature.

Hsing-yen He <hammerandsickle>
Fri 11 Sep 2015 02:15:37 PM UTC, comment #15:

Hi Steve,

I know these features performanced well in Firefox. But I just think about “locl” script tag when I found this issue, I don’t think more question about other OT features. But I feel relieved that LibreOffice for Linux can process them.

Hsing-yen He <hammerandsickle>
Fri 11 Sep 2015 01:35:43 PM UTC, comment #14:

Hi Hsing-yen,

First, these features are enabled according to text language, not by explicitly turning on font features. That is, if an application has the ability to specify the language of the text, then the feature is enabled. Mozilla and IE can do this. LaTeX with fontspec can do it. Most Word processors (last time I checked a few years ago) fail -- although they have a notion of text language, they fail to pass the information down to the font software layer.

I guess the stacked accents still aren't working the way I intended. The Vietnamese style should only appear for Vietnamese text.

What did you expect to happen with Serbian Cyrillic letters? How did you indicate that the languge of the text was Serbian? Have you read the USAGE file from the installation?

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Fri 11 Sep 2015 10:51:35 AM UTC, comment #13:

Well, I found Vietnamese-specific glyphs displaying directly in WPS Office for mobile, without any OpenType feature changes. In fact this app have no way of changing OT feature.

For this test, I rooted my phone, put font files into /system/fonts folder, installed WPS Office, and copied letters from an Wikipedia article. BTW I have also tested Cyrillic letters, no Serbian-specific glyphs appeared.

(file #34858)

Hsing-yen He <hammerandsickle>
Sat 29 Aug 2015 08:11:25 PM UTC, comment #12:

Hi Steve,

In the 20150829 build these appears better than past builds. Thanks very much.

Hsing-yen He <hammerandsickle>
Sun 23 Aug 2015 12:13:06 PM UTC, comment #11:

Hi Steve,

I understand the means of Vietnamese variant now, but I think these variants should also available in lowercase, and other faces of FreeSerif. Because there is a special typesetting for Hanyu Pinyin, this is very usual in some Chinese dictionaties and textbooks published from PRC; in this case ê (non-consonant ei) have four tones: ê̄, ế, ê̌, ề, all tone marks are just placed at the top of circumflex, they're different from Vietnamese typesetting. You can found them in Xiandai Hanyu Cidian (现代汉语词典 / 現代漢語詞典), Xinhua Zidian (新华字典 / 新華字典) etc.

Hsing-yen He <hammerandsickle>
Sun 23 Aug 2015 10:22:32 AM UTC, comment #10:

Hi Hsing-yen,

I see an inconsistency now.

For text marked as Vietnamese, circumflex-acute is consistent across faces, with the acute to the side of the circumflex.

For text marked as English,
in serif normal, the acute is placed above the circumflex, while in all other styles, it is placed beside the circumflex, as is proper for Vietnamese. in sans, in all styles the acute is placed above the circumflex. In mono. in all styles the acute is placed beside the circumflex.

I can see the table responsible in serif normal... but I forget why it is there, what I was thinking -- and what it does seems extreme (it decomposes all characters with stacked accents). This may take time to sort out.

However, this is about Latin letters, so it really doesn't belong in this report. And I am undecided as to what is "correct" here.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 22 Aug 2015 08:15:57 PM UTC, comment #9:

Good to hear, Alexey!

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Sat 22 Aug 2015 09:36:03 AM UTC, comment #8:

Доброго времени суток. Я поподробнее потестировал последние шрифты в кириллице и сделал таблички. Последние версии в MSO2003 у меня творят чудеса. Спасибо.

Alexey M-ov <teorks>
Fri 21 Aug 2015 07:04:58 AM UTC, comment #7:

Hi Hsing-yen,

you still haven't explained to me what you think is incorrect.
On my system, to my eyes, these combinations look correct.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Fri 21 Aug 2015 12:15:21 AM UTC, comment #6:

Well, I'm using the latest build, in FreeSerif Regular, I see the different Ấ and Ế after adding Vienamese markup, but other faces doesn't have this variant.

Hsing-yen He <hammerandsickle>
Thu 20 Aug 2015 08:44:51 PM UTC, comment #5:


What do you mean by "incorrect"? This is term is relative to some standard. Which do you mean?

For what application, in which language, do you use Ấ and Ế?

I hope you don't mean Vietnamese, because we discussed this before. In the HTML, simply put a
in the tag, and these accents will be positioned in a conventional way for that language.

The ä́ seems nearly perfect in my copy (see FS_BI.png). Are you sure you're using the latest build?

(file #34685)

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Thu 20 Aug 2015 05:20:36 PM UTC, comment #4:

Hi Steve,

As my test with file#34648, Ấ and Ế looks incorrect in FreeSerif Regular, FreeSans Regular, FreeSans Oblique, FreeSans Bold and FreeSans Bold Oblique; ä́ looks incorrect in FreeSerif Bold Italic, FreeSans Bold and FreeSans Bold Oblique; ã́ looks incorrect FreeSerif Bold Italic, FreeSans Bold, FreeSans Oblique, FreeSans Bold Oblique; ј̄ works badly in all faces of FreeSans.


The Z with tail (or horizontal stroke) was already existed in FreeSerif, they are variant of З (Zemlja/Ze) and Ѕ (Dzělo/Dze), they're encoded in U+A640—U+A643. These letters are used very commonly in Church Slavonic Wikipedia, but they were marked italic in some articles and thus become mojibake sometimes. I found some combining diacritical marks as you mentioned are also avalible for these letters.

Hsing-yen He <hammerandsickle>
Sat 15 Aug 2015 04:40:35 PM UTC, comment #3:

Я надеюсь, что правильно понял ваши мысли.

На счет «ꙏ» и «j with macron» очень не беспокойтесь, так как тот набор символов не критичен: это я экспериментировал на коленке с символами разных блоков. Рассчитывал на универсальное решение.
У меня есть большой файл для подбора и составления символов, с которыми приходилось сталкиваться, и которые мне реально приходилось использовать и воспроизводить по источникам (если интересно, выложу).

Насчет стилей, я не претендую, чтобы что-то абсолютизировать в конкретных положениях: это тоже больше мой беглый эксперимент. Уже думаю, что для острых и широких диакритик могут понадобиться отдельные якоря (рисунок 1 - Paint).
И работа у меня несколько специфическая, потому что представление в словаре (требующее введения некоторых специальных обозначений) и живой текст – иногда немного разные вещи.
О представлении в печатных текстах и доказательствах типа Times для некоторых символов, думаю, бессмысленно говорить, так как в гражданский шрифт (https://ru.wikipedia.org/wiki/Гражданский_шрифт) они изначально уже не попали, а существовали только в уставах (http://azbyka.ru/otechnik/Serafim_Slobodskoj/zakon-bozhij/42). (Уставной шрифт я пытался использовать только в самом начале, но отказался из-за неполноты образцов и прочих трудностей.) Так что, можно сказать, что история таких символов в новых шрифтах делается прямо сейчас. И некоторые уже существующие решения в них мне кажутся нелогичными, например «Уу – Ꙋꙋ (сравни Ɣɣ, Ȣȣ)» как будто из разных шрифтов.
И кое-что мне, конечно, хочется настроить для конкретной работы, что в общем шрифте может быть даже лишним.
1) Ну, например, встречалась мне такая проблема: литовский (рисунок 2). Для этого нужно, наверное, создать специальную диакритику.

2)Для мягкости согласных у меня принято обозначение как здесь: (рисунок 3) – то есть г̑, к̑, л̑, н̑, р̑, в заглавных словах – Г̑, К̑, Л̑, Н̑, Р̑ (а не к҄, К҄), но то же самое теоретически может понадобиться и для других согласных.

3) Кириллического í (ударный і: рисунок 4) пока в шрифтах не встречал – использую латинский знак.

4) Бывают проблемы с болгарским ударением: ъ̀ (ь̀), ъ́ (ь́) https://books.google.ru/books?ei=NNvVT8LHPOj24QTezPiIAw&hl=ru&id=VudtAAAAMAAJ&dq=%D0%B2%D1%8A%D0%BB%D0%BA%D0%B0 и сербским а̑, е̑, и̑, о̑, у̑ (ȃ, ȇ, ȋ, ȏ, ȗ), а̏, е̏, и̏, о̏, у̏ (ȁ, ȅ, ȉ, ȍ, ȕ), а̄, е̄, ӣ, о̄, ӯ (ā, ē, ī, ō, ū) – (рисунок 5). Встречается еще такой момент: Вои̍слов (https://books.google.ru/books?id=e5BiAAAAMAAJ&q=%D0%92%D0%BE%D0%B8%D1%81%D0%BB%D0%BE%D0%B2&dq=%D0%92%D0%BE%D0%B8%D1%81%D0%BB%D0%BE%D0%B2&hl=ru&sa=X&ved=0CCIQ6AEwAWoVChMI09ryzrurxwIVoUlyCh2LMAjT).

4) Для описания азбуки могут понадобиться различные вариации начертаний (рисунок 6).

Но это все вам, может быть, не пригодится.

Alexey M-ov <teorks>
Sat 15 Aug 2015 04:29:28 PM UTC, comment #2:


I just posted a build of the font (same place as before) which addresses some of the problems you pointed out.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Fri 14 Aug 2015 08:33:12 PM UTC, comment #1:

Hi, Alexey,

Thank you for all these materials. I will act on some of it.

I'm a little worried that you are spending a lot of time trying to fix things, before we discuss what you think are wrong with them. We often require documentary evidence to make changes. Please consider, you are proposing to change things that other people put a lot of time into.

First, please be aware that it is immaterial to us what other modern computer fonts do. We are interested in printed material that pre-dates computers.

Second, beware of word processors: They are terrible at typesetting and modern font technology generally. They have ceased to be a primary target for our project. Firefox is a very good environment for testing font technology.

Third, there probably is more going on in the font than you know. The placement of accents, particularly, depends sometimes on the language being used.

I'll attach an HTML version of your test.

I see (perhaps among other things):

  • Accents positioned on a couple of high letters ꙏ, ꙓ́: Frankly, I didn't know where to put them. I will simply make this change -- it's easy enough.
  • j with macron. I will need to know the application for such a thing, and evidence that the macron really goes over the j rather than replacing its dot. This is a difficult issue.
  • positioning of accents: you seem to prefer the point of the accent to be centered on the letter. This is a matter of style, but I can tell you, those accents were carefully positioned to produce an "optical" centering, common in Western typefaces. I will need evidence that optical centering is not used in Cyrillic.

broad Omegha
This letter appears in a lot of styles. One can be seen in the Unicode code charts
The form currently in FreeSerif was drawn specifically to go well with the other glyphs in the font.

To take your form, we would need visual evidence that in printed typefaces (again, we are uninterested in computer fonts) of styles similar to Times, your form of this letter is preferred.

We are very interested in your input. Persistence will be rewarded.

Steve White <Stevan_White>
Project AdministratorIn charge of this item.
Fri 14 Aug 2015 02:31:01 PM UTC, original submission:

Сверху на скрине проверки (и .doc, в архиве) результат последней, присланной мне, версии. Красным выделены моменты, к которым могут быть вопросы. Внизу - результат исправления ранее найденной мной версии. Красным выделено неисправленное, зеленым - что удалось исправить.
Больше всего интересует отображение диакритик при прописных буквах (особенно кириллических) и прочих знаках, которые уже имеют другие диакритики. Насчет вьетнамских символов, не принципиально -- просто не известно, какие комбинации понадобятся в работе еще и с древними языками. Главное, чтобы якоря работали во всех блоках (с кириллицей, латиницей, греческим).
И вопрос: как можно вообще исправлять такие проблемы в шрифтах? Возможно, придется исправлять еще и другие шрифты для словаря, но я в этом не специалист. Спасибо.

Alexey M-ov <teorks>


(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:

Attached Files
file #34699:  Pictures.zip added by hammerandsickle (1MiB - application/zip)
file #34697:  вид символов(кир.).doc added by teorks (39KiB - application/msword)
file #34694:  testfreefont-20150815 - Serif.rar added by teorks (2MiB - application/octet-stream)
file #34695:  kern.rar added by teorks (614KiB - application/octet-stream)
file #34696:  testfreefont-20150815 - Sans.rar added by teorks (2MiB - application/octet-stream)
file #34685:  FS_BI.png added by Stevan_White (104KiB - image/png)
file #34652:  А-символы.doc added by teorks (230KiB - application/msword)
file #34651:  demimg.png added by teorks (68KiB - image/png)
file #34645:  gnu-freefont_freeserif(my original).rar added by teorks (1MiB - application/octet-stream)
file #34644:  my corrected fonts.rar added by teorks (1MiB - application/octet-stream)


Depends on the following items: None found

Items that depend on this one: None found


Carbon-Copy List
  • -email is unavailable- added by hammerandsickle (Updated the item)
  • -email is unavailable- added by Stevan_White (Posted a comment)
  • -email is unavailable- added by teorks (Submitted the item)

    Do you think this task is very important?
    If so, you can add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.




    Follow 16 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2015-09-11 hammerandsickle Attached File- => Added Screenshot_2015-09-09-19-19-40.png, #34858
    2015-08-23 hammerandsickle Attached File- => Added Pictures.zip, #34699
    2015-08-22 teorks Attached File- => Added вид символов(кир.).doc, #34697
    2015-08-22 teorks Attached File- => Added testfreefont-20150815 - Serif.rar, #34694
        Attached File- => Added kern.rar, #34695
        Attached File- => Added testfreefont-20150815 - Sans.rar, #34696
    2015-08-20 Stevan_White Attached File- => Added FS_BI.png, #34685
    2015-08-20 hammerandsickle Attached File- => Added Supplement for file#34651.png, #34683
    2015-08-15 teorks Attached File- => Added А-символы.doc, #34652
    2015-08-15 teorks Attached File- => Added demimg.png, #34651
    2015-08-14 Stevan_White Attached File- => Added test-accents.html, #34648
    2015-08-14 Stevan_White Categoryoverall font problem => character range
        StatusNone => Proceeding
        Assigned toNone => Stevan_White
    2015-08-14 teorks Attached File- => Added gnu-freefont_freeserif(my original).rar, #34645
    2015-08-14 teorks Attached File- => Added my corrected fonts.rar, #34644

    Back to the top

    Powered by Savane 3.3