On Wed, May 28, 2025 at 09:33:39PM +0100, Gavin Smith wrote: > > > What I would like to change, as a preliminary step before trying to > > > find better ways of reporting on problems with the node structure, > > > is to completely separate: > > > > > > * Structure checking and warning code > > > * Code altering the structure. > > > > Looks good to me. The codes are not that dependent, it just reuse a > > loop. As it is designed now, the checks must be after determining menu > > directions, as you remarked, but there are no dependency the other ay > > around as far as I can tell. > > I've got the change at the bottom of this mail. The structure checking > comes before the structure modification, otherwise one of the warnings > is not made. I've moved structure checking code, copied the main loop > over the nodes, then combined conditional blocks where possible. > > ... > > In the next two or three days, I'll try to do the corresponding change > in the C code. Then I'll look at splitting out the structure-checking > code into its own function...
I've done this. I noticed that complete_node_tree_with_menus it was not run from C/texi2any.c, but some of this looks quite different to perl/texi2any.pl so presumably it is on purpose. > ... before looking at how this code could be > reformed or rewritten. I haven't done this yet.
