In the sequence latin \u25CC\u103A \u25CC\u103A, the first U+25CC is considered to be latin script, and this results in the following U+103A being rendered twice. The second U+25CC inherits its script from the Myanmar preceding it and the second U+103A renders fine.
test document
Logged In: YES
user_id=638379
Originator: YES
File Added: bug_25cc_103a.pdf
PDF showing the problem
Logged In: YES
user_id=638379
Originator: YES
The bug also seems to manifest itself when the second char is a subjoined char as in \u25CC\u1039\u1000, but again only when the \25CC is considered latin.
This bug is in two parts.
There was a bug in Padauk which meant that a dotted circle was not inserted automatically when U+103A starts a segment.
The duplicate U+103A occurs because graphite is always passed additional context information to ensure correct glyph selection. In this case the context causes the U+103A to be rendered with the U+25CC in the Latin segment since it is attached to the U+25CC. It is then displayed again in the CTL segment.
The context passed to graphite is determined by the general OOo code which code which switches between Latin and CTL segments.
OOo currently defaults to using the script of the previous "strong" script characters.
One way to fix this is to modify the OOo code to include "weak" script characters like U+25CC in the following CTL segment if the character is of type mark.