[L2Ork-dev] [PD] Upcoming pd-l2ork release teaser

Sat Aug 31 17:10:29 UTC 2013

On 08/30/2013 04:53 PM, Jonathan Wilkes wrote:
> Ok, got it up and running.  Some thoughts:
>
> Good:
> * wow, just click-dragging a single object around is _way_ faster than 
> in an ordinary canvas window!
> * nice semi-transparent scrollbars!
> * it's shocking how much nicer the anti-aliased lines.
> * font seems to look nicer than pd vanilla.  not sure why...
> * fontsize in floating point!
> * don't really understand the relationship between ptext and 
> transformation matrix.  it looks like you
> can rotate the text, but can you use the matrix to scale it, too?
> * matrix stuff wrt to relative location could greatly simplify some of 
> the c code. For example, an entire
> "Put" menu array is a series of relative locations of points or path 
> coords...

Thanks! :-) I will need to look into the whole matrix business.

>
> Bad:
> * how do I choose between straight lines and curved?

Currently there is no way to do this without editing the code (staright 
lines are commented out so it should not be too difficult to expand 
this). I wonder however if this is entirely necessary--03.connections.pd 
example suggests this may be so but I wonder if this is because it was 
originally organized to make things clear with straight lines. In other 
words, I tend to think that bezier cords will work in most cases except 
for those that have been explicitly designed to work with straight lines 
in which case some adjustment may be necessary.
> * in the patch menu, when i move among the menu titles I get some 
> flickering, whereas with the main pd window
> menu I don't get flickering (debian wheezy amd_64)
Is the menu flickering or the canvas? I am unfortunately unable to 
reproduce either over here (Ubuntu amd_64). If you can send me an 
example patch, that would be great.

> * I get about a 25% jump when click-dragging a single [clip] 
> continuously, compared to about 10% jump
> in cpu when in pd-0.44
This is likely since new canvas draws SVG elements as opposed simple 
lines, and a such can be more CPU intensive. My hope is that in the long 
run we finish/implement the OpenGL backend for tkpath in which case all 
this will be for the most part a non-issue.

pd-l2ork's redrawing logic is much more sophisticated than pd's. 
pd-l2ork's vanilla objects and 3rd party gui objects that have been 
adapted to take advantage of pd-l2ork's facilities, do not redraw unless 
necessary. They are relocated using canvas's move command.

> * it's a little strange that scrollbars inside patch look different 
> than the normal ones.
My hope is all scrollbars will eventually look like this :-)

> * scrollbars seem to scroll differently if you change the font size 
> (scrolls slower).

I think I just noticed that myself--should be an easy fix (hopefully!) :-)
> * Now that the scrollbars don't look two decades old, I feel like the 
> cursor should change to a normal
> arrow when they have the focus.
This may be a bit tricky due to the way canvas hijacks mouse events... 
Let me see what I can do...

> * If I duplicate [clip] about twenty times, dragging a selection 
> rectangle gets sluggish and
> moving all the [clip]s gets sluggish, too.  Sluggish = obvious 
> latency, and slower than pd vanilla -0.44
> which doesn't have dragging latency even with 40 [clip]s.
This is purely CPU overhead of SVG over conventional canvas. Once we go 
OpenGL route this will be seen only as a momentary dip in performance.

> * tried tooltips.  Hovering over left inlet of "clip":
> unknown option "-outline"
> unknown option "-outline"
>     while executing
> "$tkcanvas itemcget [lindex $tags 0] -outline"
>     (procedure "pdtk_canvas_enteritem_gettags" line 23)
>     invoked from within
> "pdtk_canvas_enteritem_gettags .x1e58d50.c 277 135 .x1e58d50.t1e58010i0"
This was fixed last night. Thanks for reporting it!

>
> Anyway, very cool so far!
>
> -Jonathan
>
>
> ------------------------------------------------------------------------
> *From:* Ivica Ico Bukvic <ico.bukvic at gmail.com>
> *To:* Jonathan Wilkes <jancsika at yahoo.com>; l2ork-dev 
> <l2ork-dev at disis.music.vt.edu>
> *Cc:* András Murányi <muranyia at gmail.com>
> *Sent:* Friday, August 30, 2013 12:23 AM
> *Subject:* Re: [PD] Upcoming pd-l2ork release teaser
>
> Guys, I think I figured out why there was a problem building tkpath. Try
> latest git commit--it appears that my .gitignore setting prevented me
> from committing some of the tkpath's files and hence the hiccup. Latest
> version also fixes lingering display issues with tkpath, route
> regression (as reported by Andras), as well as a few other cosmetic fixes.
>
> Try out the new scrolling algorithm--you can scroll with scrollbars (as
> you could before), they should resize automatically as you change the
> window size and its contents, and perhaps most importantly, now you can
> use middle click anywhere on the canvas to "drag" the canvas as if you
> were to hold paper and drag it around.
>
> HTH
>
>
>
>
> _______________________________________________
> L2Ork-dev mailing list
> L2Ork-dev at disis.music.vt.edu
> http://disis.music.vt.edu/listinfo/l2ork-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://disis.music.vt.edu/pipermail/l2ork-dev/attachments/20130831/2e7004d3/attachment.html>