Thanks Gary, for review this.<br>Would be great if we can push it in 0.94.<br><br><div class="gmail_quote"><br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> <div class="im"> > In the first test, I used alpha and scale cairo operations to modify to render only one time the icon<br> > you can see the zoom effect at the start and a little improvement in startup time:<br> ><br> ......<br> </div>Fab! That's a good 2 or 3 seconds shaved off activity start time, a very important metric, even given we loose the original pulse design (i.e. pulsing between grey+white & user colours provides better contrast; the monochrome to colour also suggests an activity bundle starting up and becoming a users own activity).<br> <br> Are these two test machines XO-1, or XO-1.5s?<br> <br></blockquote><div><br>This is in XO 1.5, In XO-1 the difference is bigger.<br> </div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> Regarding the zoom effect, it's great it is finally showing up ;-) but I wonder if the timing is different, it feels very slow/smooth/linear, is it using the same zoom maths as on previous Sugars? I'll need to have a tinker to see/check how it used to be.<br> </blockquote><div><br>Is using the same math. <br> </div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> <div class="im"><br>><br> > I did another try using two buffers, with the different colors, and changing the alpha<br> > to obtain a similar effect. The code is a lot more complex, and is a little slower obviously.<br> ><br> > <a href="http://www.youtube.com/watch?v=9rLaE0TKZXw" target="_blank">http://www.youtube.com/watch?v=9rLaE0TKZXw</a><br> ><br> > <a href="http://www.youtube.com/watch?v=-Q2RliMTLRM" target="_blank">http://www.youtube.com/watch?v=-Q2RliMTLRM</a><br> ><br> > <a href="http://www.youtube.com/watch?v=SHBCSnrGsjA" target="_blank">http://www.youtube.com/watch?v=SHBCSnrGsjA</a><br> <br> </div>I do find these more visible as the grey/white stage of the pulse is guaranteed to provide some contrast even if the user has chosen a fill/stroke combination with low contrast. However, the cost seems rather high with only about 1sec launch time improvement and twice as much code to maintain.<br> <br></blockquote><div><br>Yes. That is the reason because I need your opinion about if the first approach is enough or if we need the second one.<br> </div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> Quick question about your second version (I couldn't quite grok my way through the patch). Are you 1) fading both the grey and the colour versions of the image so that as one becomes opaque, the other becomes transparent; or 2) drawing one in the background always as an opaque fill, and then just pulsing the other over the top from opaque to transparent? In my code (non Sugar) I find the latter choice keeps code to a minimum, is more efficient, and visually looks cleaner (avoids the white background leaking through during part of the cycle).<br> <br></blockquote><div><br>I am doing what you describe in 1)<br>I can try what you propose in 2) . Looks like is not a big change in the code,<br><br> </div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> Here's a quick video if I didn't describe the above effect well enough (middle is alpha fading both images, right is the simple case of a single fade in/out for comparison, and left is using the monochrome image as an opaque background fill while drawing the colour image over the top with an alpha puse):<br> <br> <a href="http://youtu.be/LLsqjGWBMfQ" target="_blank">http://youtu.be/LLsqjGWBMfQ</a><br> <br></blockquote><div><br>Thanks. Is very clear.<br><br>Gonzalo<br><br><br></div></div>