Age | Commit message (Collapse) | Author | Files | Lines |
|
Word list authored by: sorpaas
Sources:
lo gimste jo'u lo ma'oste (http://guskant.github.io/lojbo/gismu-cmavo.html)
N-grams of Lojban corpus (https://mw.lojban.org/papri/N-grams_of_Lojban_corpus)
|
|
Word list authored by: Engelberg, ProkhorZ
Sources:
Baza Radikaro Oficiala
Reta Vortaro (http://www.reta-vortaro.de/revo/)
Esperanto Panorama - Esperanto-English Dictionary (http://www.esperanto-panorama.net/vortaro/eoen.htm)
ESPDIC - Paul Denisowski (http://www.denisowski.org/Esperanto/ESPDIC/espdic.txt)
|
|
|
|
|
|
|
|
and a test to go with it
Remember to run the test when changing word lists, or simplewallet
will throw uncaught if that word list is used.
|
|
|
|
If a checksum word is present, language detection would use
just the word prefixes. However, a set of word prefixes may
be found in more than one language, and so the wrong language
may be found first, which could then fail the checksum, since
the check may be done with a different unique prefix length
from the one it was created from.
We now make a checksum test when we we detect a language from
prefixes only, to make sure we have the correct one.
|
|
Ain't nobody got time for link/cmake skullduggery.
This reverts commit fff238ec94ac6d45fc18c315d7bc590ddfaad63d.
|
|
Useful for debugging users' logs
|
|
Some word triplets, such as "mugged names nail", are not valid
results from any 32 bit value. If used to decode a 32 bit value,
the result will therefore encode to a different word triplet.
Fix this by using random words converted from an actual random
bitstring, ensuring we always get valid triplets.
|
|
|
|
|
|
|
|
|