You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@openoffice.apache.org by bu...@apache.org on 2012/08/10 17:43:47 UTC

[Bug 120530] in draw:transform value rotate the angles are treated as rad but ODF specifies deg

https://issues.apache.org/ooo/show_bug.cgi?id=120530

Armin Le Grand <Ar...@me.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |Armin.Le.Grand@me.com

--- Comment #1 from Armin Le Grand <Ar...@me.com> ---
ALG: The description is wrong, indeed rad is used (0 .. 2pi) which is related
closest to draw:transform in it's sense as a transformation matrix. the
'rotate(..)' statement is used for better readability, only.

More bad: Just checked, rect with 45deg in UI (thus, rotated left) saved, gives
'rotate (0.7853981633973)'. When decomposing the matrix (which is not written)
you would get -0.78 which would be mathematically positive in the screen
coordinate system.

In short: the API and ODF error is by purpose used here, too, see
xmloff/source/draw/xexptra.cxx lines 664 ff (#i78696#).

Probably for consistency (can't remember completely) to be able to use the API
object rotation directly. See xmloff/source/draw/shapeexport2.cxx ln 196 where
the value coming from a matrix decmopose (and being correct, mathematically
positive) is mirrored before export.

Thus, 'rotate(..)' from draw:transform of 19.228 is:
- in rad
- negative oriented

Where the angle contained evtl. in 'matrix(..)' from draw:transform of 19.228
is:
- in the units you get when decomposing it
- positive oriented

-- 
You are receiving this mail because:
You are the assignee for the bug.