[Nauty] Group-related issues

[Anton Voropaev]

It might be convenient to call `userlevelproc' for the order-zero
graph in `nauty'. Then `grouplevelproc' could properly update
the group structure, which currently remains untouched.

The `cut loose' mechanism in `naugroup.c' can corrupt the coset
extension of a group structure, leaving it reused until the next
`cut loose' request or reallocation.

The attachment contains illustrations and fixes.


Typos.


nug27.pdf (page[[, paragraph], line])

18, 1: is (are?)
27, 4, 2: defined above (does not seem to be)
29, 4: i1 i2 ... il (il i(l+1) ... ik)
38, -4: smallest (positive?)
41: SETWORDSNEEDED twice
43, -4: verion
65: followed by ./runalltests (redundant)
81, -3: not found
81, -1: mismatch
92: overfull
103: overfull


README

29: have have


nauty-h.in

289: the the


nauty.c

177: the the
343: stats (stats_arg?)
1176: automophism


labelg.c

3: [sg | C#W#] (hyphens?, z?)


genbg.c

28: subset of another vertex ('s neighbourhood?)


genquarticg.c

1115: the the



Anton Voropaev
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fix.tar.gz
Type: application/x-gzip
Size: 4047 bytes
Desc: not available
URL: <http://mailman.anu.edu.au/pipermail/nauty/attachments/20181220/aa8a9e72/attachment.bin>