[Gtkradiant] [Bug 663] Gensurf crash

gtkradiant@zerowing.idsoftware.com gtkradiant@zerowing.idsoftware.com
Sun, 10 Nov 2002 10:57:15 -0600


------- Additional Comments From ttimo@idsoftware.com  2002-11-10 10:57 -------
TTimo: the entity ends up in map when you commit it, it gets built at that time
TTimo: so I'd say it is 'expected behaviour'
RR2DO2: but for some reason both bobtoolz and gensurf did this:

create epairs
create brushes
add entity handle to map

instead of 

create epairs
add entity to map
create brushes in it
TTimo: well, the #1 solution always worked fine till recently 
TTimo: why ?
TTimo: someone changed the plugin API and causes rebuilds when brushes are
commited to ent I suppose ?
TTimo: (not API actually, implementation)
RR2DO2: yeah I think I did that cause of a problem I encountered with bobtoolz
(it's in my changelog)
RR2DO2: ah I know
RR2DO2: it was related to rebuilding brushes
RR2DO2: I would delete a brush
RR2DO2: and replace it with a new one
RR2DO2: but because the entity handle never gets comitted to the map, after all,
it;s already there
RR2DO2: the brush I just added never appeared
TTimo: so when you commit a brush to an entity, you should check if the entity
is already in map, or being built
TTimo: and decide wether to rebuild or not 
RR2DO2: *nods*

well we could delete the three offending lines in the api for now, and revert
the patch on gensurf. I can then track down the issue closer over here (well at
the office tomorrow) and see if I can work around it
TTimo: sounds good to me
RR2DO2: it's in CommitBrushHandleToEntity,

      Brush_AddToList (pb, &active_brushes);

those three lines
TTimo: k