We were erroneously using canvas_getrootfor instead of glist_getcanvas to
find the canvas on which the nested gop was being displayed.

Also added a consistency check to turn such a crasher into a runtime bug
in the future.

First checkpoint in implementing custom scrollbars and improving the scrollbar behavior. This is a WIP that requires nw.js 0.46.2.

Finished second stage of the scrollbar implementation and added scrolling with middle click. All scrollbars are now operational and to the best of my knowledge accurate. The only things remaining are: getscroll needs to be called on deselect to recalculate when the handles are erased on objects that have them (we should get rid of these and focus on bboxes instead), responding to maximizing and restoring, and addressing a weird bug where in 0.46.2 patches are saved with larger window sizes under certain conditions.

Reverted scrolling by grabbing due to problematic relationship between scrollBy and the zoom factor

Continuing work on the scrollbars. The only known thing now remaining is ensuring that the patch window is properly recorded when saving it (under certain conditions it spits out completely wrong values

Removed debug printouts

Finished the scrollbars. The only thing remaining is for someone who has a better understanding of the pd_canvas.js and how the events are captured, to refactor my registration of events that ensure scrollbars are clickable and the onresize event. Currently, they are embedded inside the pd_canvas.js under the body tag where you will find both onscroll and onresize events.

Overhaul of the new scrollbars to further improve their behavior, enable as perfect of a fit of the plots both in gop and on toplevel, and dynamic updating of the plots in toplevel when resized. Reworked dynamic event adding/removal for the scrollbars and removed onscroll from the html document and moved it into the pd_canvas.js. Still need to add a delay for the backend callback to update the scrollbars on the toplevel plot/array and thereby remove scrollbar flicker.

Finished implementing the scrollbar callback and addressed popup not appearing in the right place (likely due to upgrade to 0.46.2)

Added names of arrays to their toplevel subpatch window title. Multiple arrays are listed in the order they have been added.

Another reworking of scrollbars to and array sizing to capture all possible scenarios. Some refactoring to make use of the canvas_hasarray. Remaining known issues pertaining to the plots are:
*array toplevel windows are not restored in position they were saved
*last element on the plots is not clickable
*make bezier plot look like bezier (may need to leverage curve_path)

Jonathan Wilkes authored 4 years ago and Ivica Bukvic committed 4 years ago

This was initially part of an effort to prevent double binding to scalars
for certain event callbacks. By unvis'ing all scalars in the graph this
would ensure the symbol would get unbound.

However, we're already doing that in canvas_map. So it appears this extra
unvis'ing in graph_vis was unnecessary.

However, it looks like the scalar in a gop will get unvis'd twice when
destroying the gop canvas. No idea why that is, but if there's a regression
perhaps one of the few people using scalar events will report a regression :)

Jonathan Wilkes authored 4 years ago and Ivica Bukvic committed 4 years ago

* give new cord the same color as selected line (except for vanilla/ext presets)
* make broken object stand out more
* make comment dash sizes match l2ork 1.0
* change xlet hover animation to match new xlet style

This was initially part of an effort to prevent double binding to scalars
for certain event callbacks. By unvis'ing all scalars in the graph this
would ensure the symbol would get unbound.

However, we're already doing that in canvas_map. So it appears this extra
unvis'ing in graph_vis was unnecessary.

However, it looks like the scalar in a gop will get unvis'd twice when
destroying the gop canvas. No idea why that is, but if there's a regression
perhaps one of the few people using scalar events will report a regression :)

Fixed minimum array gop size on the parent window due to the garray text size which was before spilling over the minimum box size and not maintaining proper margins.

Disabled moving canvas with arrows in non-edit mode since we may want to use those keys through the [key] object. Later we can expose scroll ability to use key object to drive scrolling...

Additional commenting to make it easier to catch-up with the Wagnerian function dealing with the key input sent by the front-end. Disabled check that prevented exiting from the edit mode when creating a new object via a keyboard shortcut on a canvas in a non-edit mode. Upon releasing CTRL, the edit mode was not initiated.

Various cosmetic improvements that deal primarily with proper highlighting of object borders, with particular focus on comments. See comments in the code for more info.

Proper fix for the nlet highlighting. Now allows for loose aiming and always starts the patch cord at the center of the outlet. Inlets from which we currently cannot connect from are grey when highlighted, while ones you can initiate connection from are orange (both colors are themeable inside the css). Also, if trying to connect an impossible connection (signal to non-signal), when hovering over the non-signal inlet during the connection cycle, such an inlet will be also highlighted gray (grey? whatever...). Lastly, updated the highlighting animation--I am really not particularly interested in animation as it tends to slow down the interaction with the patch, so the animation is much shorter and more minimal. Potential value is leaving it as this to provide visibility of the nlet being highlighted when we are already over it with the cursor, which will otherwise occlude the nlet altogether. Another consideration with the currently committed highlight is that it is hollow inside which looks IMO distracting. Curved edges is also distracting since no other aspect of the UI (save for the customized data structures created by the user) is not curved.

Long and painful journey to make all plots fit into a GOP graph is (hopefully) complete. One thing I did not test is having multiple plots in a single GOP, even though it should work ok as long as each scalar group is treated as separate.

Improved variable wording and removed extraneous printouts.