On Wed, Nov 18, 2015 at 10:36:54PM +0100, Jean-Marc Lasgouttes wrote:
> Le 18/11/15 22:27, Scott Kostyshak a écrit :
> > From what I understand, you use a static function as opposed to a static
> >variable because it needs to be constantly recomputed in case, e.g.
> >there is a zoom. First, I do not understand why the value changes for me
> >if there is a zoom (which it does seem to do in testing) whereas the
> >value was fixed before. Since I do not have a HiDPI I did not expect a
> >difference.
> 
> Well, you can expect that the spacing is, say, half of the width of a 'm'.
> So if you zoom in, the size increases.

I just did not expect this change from your patch's description since
before, Inset::TEXT_TO_INSET_OFFSET was fixed. I thought that there
would only be improvements for HiDPI.

> >Second, if it does need to be recomputed, would it be more efficient to
> >only recompute it when triggered by certain events (e.g. a zoom) instead
> >of recomputing it every time?
> 
> I thought about it, but it remains to be proved that computing the value is
> expensive. It is just a few multiplications.

I see. Makes sense.

> Nothing compared to painting or
> measuring text

or checking whether previews need updating, as I have learned.

Scott

Attachment: signature.asc
Description: PGP signature

Reply via email to