[Noto Serif] Incorrect positioning of Cyrillic diacritics U+0483..U+0488 over Cyrillic letters in "Noto Serif" (correct in "Noto Sans")

Specify the font name in title with a short description of the bug.
Please report any issue related to Noto fonts [here](https://notofonts.github.io/reporter.html). (not relevant, this is for script-specific fonts, not the base "Noto Sans" or "Noto Serif" that are not selectable there).
Report any issue with Google Icon/Symbols [here](https://github.com/google/material-design-icons) (not relevant).

**Describe the bug**
In "Noto Serif", the Cyrillic diacritics U+0483..U+0488 are incorrectly positioned to the right of Cyrillic letters after which they are combined, instead of being centered on these letters. The positioning is correct in "Noto Sans".

These combining diacritics, not encoded in the generic Diacritics bloc (titlos above, or surrounding accents) are mostly used for noting traditional Cyrillic numbers, still used in Church Slavonic, and that were used in old Russian, where Cyrillic letters take their numerical value according to the Greek alphabetic order from which Cyrillic letter forms were inherited, rather than their modern Cyrillic alphabetic order. (These Cyrillic diacritics for numbers have some equivalents in the Greek alphabet for their numerals, using either a prepended lower numerical mark, or an appended upper numerical mark, or a combining psili pneumata, and sometimes the Cyrillic diacritics for some more recent extension). A few other additional Cyrillic diacritics for numerals are present in Cyrillic Extended blocks.

**To Reproduce**
Just an example (there are many combinations for each one of the 27 Cyrillic base letters and every numeral diacritic marking a multiplier by a power of ten, instead of using the positional decimal system with a modern digit zero):
- U+0410 U+0488 (complete numeral for one million)

**Expected behavior**

**Screenshots**
In examples below, the diacritics are rendered with the Cyrillic small letter o, for clarity, but this occurs as well with every other base Cyrillic (capital or small) letters.

* With "Noto Serif" (incorrect, see the blue cells):
* With "Noto Sans" (correct, see the blue cells):

Note that:
* Some Cyrillic diacritics also exist in two half forms, notably the "combining titlo above" to extend it over multiple letters, each one with a combining part, possibly joining these two half parts by adding a joining "combining macron above" to each intermediate letter; these half forms encoded outside this Unicode range, are not rendered here, but are also affected by this bug in "Noto Serif". However their use is much less frequent, except in some medieval or litturgic Slavonic texts, or old inscriptions and coins, to mark dates, amounts, honorific ruler names, and numbering chapters/pages/articles of old legal texts.
* Cyrillic numerals using these diacritics were historically used only with older Cyrillic letter forms (in ISO 15924: "Cyrs" instead of "Cyrl"), notably for the Cyrillic letters A and BE representing the values 1 and 2, both styles being unified in code points assigned to the UCS by the Unicode and ISO/IEC 10646 standards. However there modern texts in Church Slavonic and Russian using modern letter forms still used today. Both styles for modern Slavic "Cyrl" and traditional Slavonic "Cyrs" should be supported in "Noto Sans" and "Noto Serif" (at least for all the 27 basic Cyrillic letters used as numerals with units 1 to 9, tens 10 to 90, and hundreds 100 to 900, and according to their equivalent Greek alphabetic order; however for some basic Cyrillic numerals, there are additional alternate Cyrillic letters used in isolation or with these multiplying diacritics).
* The use of "titlos" diacritics in Cyrillic numerals was optional: they were used to disambiguate numerals from normal text, the same way as leading or trailing numeral marks were used in Greek numerals with Greek letters, or middle dot marks were used in Roman numerals with Latin letters, or leading numeral marks are used in modern Braille.