| ID: | 2655 |
| From: | Gabriel Paubert <pa...@iram.es> |
| Date: | Wed, 20 Feb 2019 08:15:43 +0100 |
| Subject: | Re: [pcb-rnd] poly clipping performance optimization - experimental |
| in-reply-to: | 2631 from ge...@igor2.repo.hu |
| replies: | 2658 from ge...@igor2.repo.hu |
Hi,
On Thu, Jan 31, 2019 at 07:28:55AM +0100, gedau@igor2.repo.hu wrote:
> Hi Gabriel,
>
> in the previous mail I wrote about how the fix involves moving the
> decision about when we do the clipping one level higher, e.g. to the
> "operation on all selected objects" function.
>
> This morning I added an experimental action that can move it even higher:
> to the user level. Please test if it helps your workflow on your big
> board. I expect it could help you not needing to do drawing outside of the
> board. If it does help, we may have this included as a standard feature in
> the next cycle. If it doesn't help, I will revert the action.
>
> Steps to get the new action:
>
> 1. configure with --buildin-expfeat so that you get the plugin with
> experimental features
>
> 2. use the ClipInhibit() action with one argument (on, off or toggle) to
> change the current inhibit state. During an inhibition (on), no poly
> clipping is done on the board. This means when you draw a new polygon, it
> won't show up, when you put objects within a poly, they won't get
> clearances, etc. Tip: you may want to turn on the wireframe draw mode
> while inhibit is on (the | key) to make new objects visible under
> polygons.
>
> 3. once you are done with a lot of operations that affect large polygons,
> turn off the inhibit. This will do all the clipping at once. If it takes
> longer than a few seconds, it should draw a progress bar.
>
> 4. attached is an optional menu patch; it doesn't have a checkbox, it's
> not easy to do without a proper conf item which I didn't want to add
> before we figure whether the feature is useful at all. But it has a hotkey
> {m p i} and a menu item so it's easier to call it. It toggles clip
> inhibit on each ionvocation.
>
>
> If you have the time, please experiment with it and let me know if it
> looks useful. So far it was super cheap to implement and in expfeat it is
> not introducing any risk on our upcoming release.
These last days I have been using this experimental feature extensively,
it really speeds up thing. Now my fingers have learnt the {m p i}
sequence!
This said, when I forget to inhibit clipping, there is one function
which really is incredibly slow in many cases: undo. Maybe the clipping
should be always inhibited at the beginning of undo and restored after.
Regards,
Gabriel
Reply subtree:
2655 Re: [pcb-rnd] poly clipping performance optimization - experimental from Gabriel Paubert <pa...@iram.es>
2658 Re: [pcb-rnd] poly clipping performance optimization - experimental from ge...@igor2.repo.hu
2672 Re: [pcb-rnd] poly clipping performance optimization - experimental from ge...@igor2.repo.hu
2673 Re: [pcb-rnd] poly clipping performance optimization - experimental from Gabriel Paubert <pa...@iram.es>
2674 Re: [pcb-rnd] poly clipping performance optimization - experimental from ge...@igor2.repo.hu