add an assert for inserting an egde with the same vertices and a BMESH_TODO for bevel.

This commit is contained in:
2012-02-29 16:29:09 +00:00
parent ce64128a84
commit 6c53638863
2 changed files with 20 additions and 0 deletions

View File

@@ -94,6 +94,9 @@ void BLI_edgehash_insert(EdgeHash *eh, unsigned int v0, unsigned int v1, void *v
unsigned int hash;
EdgeEntry *e = BLI_mempool_alloc(eh->epool);
/* this helps to track down errors with bad edge data */
BLI_assert(v0 != v1);
EDGE_ORD(v0, v1); /* ensure v0 is smaller */
hash = EDGE_HASH(v0, v1) % eh->nbuckets;

View File

@@ -43,6 +43,23 @@
#include "bmesh.h"
#include "bmesh_private.h"
/* BMESH_TODO
*
* Date: 2011-11-24 06:25
* Sender: Andrew Wiggin
* Status update: I have code changes to actually make basic bevel modifier work. The things that still need to be done:
* - clean up the changes
* - get bevel by weight and bevel by angles working
* - the code uses adaptations of a couple of bmesh APIs,
* that work a little differently. for example, a join faces that doesn't just create a new face and then delete the
* original two faces and all associated loops, it extends one of the original faces to cover all the original loops
* (except for the loop on the join edge which is of course deleted). the bevel code currently requires this because it
* expects to be able to continue walking loop lists and doesn't like for loops to be deleted out from under it
* while working...
* but bmesh APIs don't do it this way because it makes it trickier to manage the interp during these operations,
* so I need to decide what to do in these cases.
*/
/* ------- Bevel code starts here -------- */
BME_TransData_Head *BME_init_transdata(int bufsize)