Closed ids1024 closed 5 years ago
Hm. Honestly, I'm not familiar enough with the format of this file to know exactly how it should be, so I don't know if there's anything wrong with my change (which changes it to match the DICTLINE.GEN.csv
from the other PR). I don't know if you're more familiar with it.
How did you get that error message?
it runs fine for me with a few test words, but results in Exception in PARSE_LINE processing regina gravi iamdudum saucia cura
when run on test/aeneid_bk4.txt
.
It looks like we have a packaging error; one of the data files is actually programmatically generated from one of the others, and this isn't captured in the Makefile. It is run through a tricky sort programme that knows about Latin's collation rules (i.e., I = J, U = V), and the results are cached in the repo. There were good reasons for this in the 1980s, but I reckon they no longer apply and that we should change this; after the change we'll have another look at this PR.
ok - build system fixed, test system modified to be less sensitive to changes in the dictionary. merged
thanks for sorting this out @ids1024
thanks - note that this triggers a problem which prevents the programme from running; it might be a bug
raised CONSTRAINT_ERROR : words_engine-list_sweep.adb:123 discriminant check failed