On Fri, Oct 7, 2011 at 9:53 AM, Jim Talbut <[email protected]> wrote:
> Hi,
>
> I've finally got my images sized correctly in both HSSF and XSSF with POI
> 3.8 beta 4 and I'd be grateful if someone would confirm the behaviour I've
> found.
>
> For HSSF the Dx and Dy values are based on fractions of a cell width/height,
> ranging from 0 to 1024/255.

quite right.

> For XSSF the Dx and Dy values are EMU values, with an upper bound that
> depends on the size of the column/row.
>

Logically yes, the upper bound depends on the size of the window, but
the format spec does not impose this restriction.

> Is this the expected behaviour for 3.8 beta 4?
> Are you expecting this behaviour to change in the release version of 3.8?
>

Yes, it is expected. It may be confusing that implementations of
ClientAnchor behave differently (fractions vs EMUs), but I don't see
how to make this differences transparent to user. Dealing with
fractions is a pain. EMUs are much easier and personally I would be
happy if both HSSF and XSSF operate in EMUs but it is not possible for
historical reasons.

> I'm quite happy with the behaviour (now I know what it is) but I couldn't
> find the XSSF behaviour documented anywhere.
> The comments in the ClientAnchor javadocs are confusing, because whilst they
> correctly imply that XSSF values are EMU based, they also imply that HSSF
> values are pixel based, which they aren't.
>

Something to improve for 3.8-FINAL. You are very much welcome to post
the corrected javadoc for ClientAnchor :)

Regards,
Yegor

> Thanks.
>
> Jim
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to