[Bug 499790] [te_IN][pango]GSUB delete key can't delete the whole char in f10

bugzilla at redhat.com bugzilla at redhat.com
Thu Oct 9 09:16:55 UTC 2014


https://bugzilla.redhat.com/show_bug.cgi?id=499790

Akira TAGOH <tagoh at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |CLOSED
         Resolution|---                         |NOTABUG
        Last Closed|                            |2014-10-09 05:16:55



--- Comment #10 from Akira TAGOH <tagoh at redhat.com> ---
After taking a look at UAX#29 and the Pango code, current behavior is following
the Unicode standard unfortunately.

Here is the details how the text segmentation is decided in Pango according to
the standard:

First, the character type for the above characters are:
U+0C15  OTHER LETTER
U+0C4D  NON SPACING MARK

Pango is looking at the grapheme boundary rule to decide if a character should
be deleted by backspace key. it's defined as following in UAX#29:

* Break at the start(GB1) and end(GB2) of text.
* Do not break between a CR and LF(GB3). Otherwise, break before(GB4) and
after(GB5) controls.
* Do not break Hangul syllable sequences(GB6,GB7,GB8).
* Do not break between regional indicator symbols(GB8a).
* Do not break before extending characters(GB9).

Only for extended grapheme clusters:
* Do not break before SpacingMarks(GB9a), or after Prepend characters(GB9b).
* Otherwise, break everywhere.

According to the table 2 in UAX#29, non-spacing mark is categorized into
Extend. so it doesn't break *before* U+0C4D. but no rules to not break *after*
Extend.

At this point, there are no standard that the software can refers the material
to prevent that behavior. maybe good to escalate this to Unicode.org then.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=7ZouYbGPAL&a=cc_unsubscribe


More information about the fonts-bugs mailing list