On Mon, Nov 02, 2020 at 12:30:18AM +0100, Patrice Dumas wrote: > On Sun, Nov 01, 2020 at 07:26:52PM +0100, Patrice Dumas wrote: > > On Sun, Nov 01, 2020 at 11:40:56AM +0000, Gavin Smith wrote: > > > On Sun, Nov 01, 2020 at 01:37:04AM +0100, Patrice Dumas wrote: > > > > > In that case, > > > > _default_node_direction() should probably be changed, to give something > > > > reasonable when there is no node associated with the section. Actually, > > > > it would probably make more sense to use sections for directions if > > > > USE_NODES is set to 0. > > > > > > I haven't investigated what happens currently but this might make sense > > > so that the section has "Next" and "Prev" links. > > > > Ok. I will probably try to do something for that case. > > Here is what I propose. We add a new variable, like > USE_NODES_DIRECTIONS. It can be set to 1, 0, or undef, the default. > If undef it follows USE_NODES.
Do you mean that it takes the same value as USE_NODES? That sounds OK. > > If USE_NODES_DIRECTIONS is true, use the same code as today, if it is > false, use the section directions instead of the nodes directions in > _default_node_direction (this function would be renamed too). > > Does it looks good? I'm OK with it if you think it is necessary. I would call it USE_NODE_DIRECTIONS instead.
