The internal code for render layers node is unchanged, since the buffer is created anyway. Also noticed that, while unavailable sockets and links to/from them are not displayed, they are still used in execution, which could lead to unexpected results. But that's a more general problem with the way (pseudo-)dynamic sockets are simply flagged with SOCK_UNAVAIL.
* Disabled possibility to keyframe render resolution, percentage and border render properties. Animating them does not makes much sense. Discussed with Sergey in IRC.
for viewport display.
It saves plenty of memory when using subsurf modifier which result is getting
converted from CCGDM to CDDM without any benefit. It also syncs behavior of
modifiers with mesh.
Need to keep an eye on constructive modifiers when in edit mode.
Discovered this when was looking into #29973.
Issue was caused by resetting face's mat_nr to zero if there's no material
map sent to ConvertCSGDescriptorsToDerivedMesh. In case of boolean modifier
we can't use such map because we can't affect on materials present in object.
So the only way which can give reasonable result is:
- Dot change mat_nr for faces from left operand (they should be fine, because
materials aren't deleting by modifier)
- For faces from right operand check if needed material exists in left operand
and if so, use it's index as new mat_nr.
- If there are materials in right operand which doesn't exist in left operand,
they'll be changed to first material from left operand.
While looking at a bug report, I found that the current GUI for the Action
Constraint actually didn't make sense, mixing up settings so that it wasn't
clear which settings corresponded to which others. This commit cleans up the
layout into a clearer two-column design to have a "from" -> "to" layout, making
all the
relationships between things clear.
For more details see http://aligorith.blogspot.com/2012/01/action-constraint-
gui-revised.html
- ui_check_but() in ui_textedit_move() is necessary because this function clips
but->drawstring to fit text entry widget and it confuses cursor movement stuff.
ui_check_but copies editstring to drawstring, so displaystring would be clipped
again in correct way.
- If the whole drawstring fits widget, no need to set button's offset.
Issue was caused by changing button state to EXIT, so there's no active button just after
applying Alt-Wheel event. Setting this state is needed to prevent button trigger cancel
callback when mouse is leaving hovered menu button.
Using the same "post activate" trick used by Tab button allows to make prevent canceling
button and makes this button active again after applying all handlers.
There's still issues with Alt-Scroll if changing active element in menu leads to interface
changes (like file format in render buttons) -- in this case button simple doesn't receive
wheel events and it's actually not connected to this issue.
Crash was because the loop walker's wire edge handling wasn't general enough (handled only wire edges that were connected to exactly one other wire edge)
- Renamed graph_jump_to_current_frame to graph_center_current_frame
which makes more sense.
- Curve view now can be locked to time cursor (Lock to Time Cursor in
Display panel or L button in curve view). Not sure if offset from
locked position will make much sense here.
- Added hotkey for solving -- Shift-S.
- Use proper poll functions for tracks curve operators.
- Darken frames outside of scene frame range in curves view.
- Implemented view all operator.
- Implemented jump to current frame operator.
Should be no functional changes, but this discovered inconsistency in structures
used for armature circle selection, not sure how harmful this were, but it's not
nice to have such inconsistencies. Anyway, it's resolved now.
