In fact there were several issues fixed (all of them regressions since bmesh merge):
- Creating navmesh crashed because creating new faces for mesh was trying to set
default values for all customdata layers in this face. This requires memory
pool created for this datablock.
Usually this pool is creating on creating datablock if there're some elements
to be stored in this block. In cases of regular primitive creating it wasn't
an issue because they doesn't create customdata layers, they only creates
geometry.
Navigation mesh creates geometry and customdata layers (CD_RECAST layer)
which used to confuse a bit custom data functions. Solved by ensuring there's
memory pool created for polygons datablock after adding new custom data layer.
Most probably it's better to be resolved on CD level (like smarter track on
changed amount of stored data and so) but prefer not to make such global changes
so close to the release.
- Toggling edit mode lead to loosing recast datalayer. Solved by adding recast
layer to bmesh mask so it'll be copied to/from edit mesh.
- Some part of code assumed raycast layer is in face datablock, some that it's in
polygon datablock. Made it to be in polygons datablock.
Kind of temporary solution to make navmesh working, probably it'll fail if one
will want to edit navmesh by hand after it was generated.
Proper way would be to ensure the whole navmesh things are using ngons.
When ORIGSPACE is added, it is now a loop data, so we need to mark tessdata as dirty too. Also fixed DM_ensure_tessface to check DM_DIRTY_TESS_CDLAYERS flag!
This fixes problems with subsurf + dynapaint + particles, however, some modifiers, like remesh, seems to create a new dm from scratch, hence loosing completely those CD layers… Note this bug already existed in 2.62, so this is not a regression.
--debug
--debug-ffmpeg
--debug-python
--debug-events
--debug-wm
This makes debug output easier to read - event debug prints would flood output too much before.
For convenience:
--debug-all turns all debug flags on (works as --debug did before).
also removed some redundant whitespace in debug prints and prefix some prints with __func__ to give some context.
Should also fix [#30266], [#29451], and partly [#30316].
Here are the changes made by this commit:
* It adds a "dirty" flag to DerivedMesh struct (for now, only DM_DIRTY_TESS_CDLAYERS, but more might be added as needed).
* It adds a new func, DM_update_tessface_data, which assumes tessfaces themselves are valid, but updates tessellated customdata from their poly/loop counter parts.
* At end of modstack, when valid tessellated faces are present in finaldm , but the cdlayers dirty flag is set, call that function (instead of recomputing the whole tessellation).
* Edits to the codes concerned (UVProject, DynamicPaint, and Subsurf modifiers).
* Also add to subsurf dm generation code the creation of a CD_POLYINDEX layer (mandatory for DM_update_tessface_data to work well, and imho all tessellated dm should have one).
Note: some pieces of old code are just #if 0’ed, will clean them later.
Issue was caused by MFACE layer adding for even DM without tessellated faces
which lead to adding new layer but with NULL data. This causes issues when layer
with faces (after extrude) was attempting to add because currently CD only checks
if layer type exists but does not check for size of the layer.
Solved by not adding MFACE layer if there's no tessellated faces.
This idea is borrowed from the multires modifier, which already
checked if the object was in sculpt mode and, if so, created the
PBVH. That check is now moved higher up the chain into
mesh_build_data(), so that it occurs for CDDerivedMesh too.
This also replaces an assert in cdDM_getPBVH for tesselated mesh faces
with a call to create them if missing.
- spelling - turns out we had tessellation spelt wrong all over.
- use \directive for doxy (not @directive)
- remove BLI_sparsemap.h - was from bmesh merge IIRC but entire file commented and not used.
These changes are to make the bmesh api more consistent and easier to learn, grouping similar functions which is convenient for autocomplete.
This uses similar convention to RNA.
* use face/loop/edge/vert as a prefix for functions.
* use 'elem' as a prefix too for functions that can take any type with a BMHeader.
* changed from camel case to underscore separated (like RNA).
In fact, we have to add a CD_WEIGHT_MCOL layer (if not yet present) when updating CD_WEIGHT_MLOOPCOL, else with modifiers modifying the topology (like subsurf), it will sigsev.
That step should probably be done at tesselation time, though…
Also simplifies DM_update_weight_mcol: no need to update CD_WEIGHT_MCOL here, as it is anyway recreated from CD_WEIGHT_MLOOPCOL at tesselation time. Only commented out code for now.
User side:
* Preview for DynamicPaint should keep the same behavior (for now). Weight preview should be somawhat quicker, though.
* Preview for WeightVG modifiers is only active in WeightPaint mode, and if the affected vgroup is the active one.
* Last active preview modifier in stack wins!
Note: that modifier preview topic is yet to be further refined, quite raw/incomplete for now.
Dev side:
* In draw code, renamed DRAW_DYNAMIC_PAINT_PREVIEW flag to DRAW_MODIFIERS_PREVIEW
* Removed use of MOD_DPAINT_PREVIEW_READY in DynamicPaint code (seems unecessary, and if it was, should be of more general scope).
* Added eModifierTypeFlag_UsesPreview to ModifierTypeFlag, for modifiers that can generate some preview data.
* Added three new modifier funcs, to handle preview modifiers in draw code / mod stack.
* For weights preview: added the generic DM_update_weight_mcol func, which can update WEIGHT_MCOL layer with either a given array of weights (currently used by DynamicPaint only), or from current active vgroup(s).
So now, draw code is fully generic (i.e. no more modifier-type checking in it). Mod stack code is generic to some extent, but will need more work.
