Closed tggo closed 1 year ago
@tggo The tests currently fail due to differences between the expected output in your snowball-data PR and what your algorithm actually produces, all involving apostrophes from a quick look. Please can you resolve this?
All languages are failing except Pascal (which passes because it only supports stemmers which can work in ISO-8859-1 currently) and Ada (which was passing due to a bug in the the make rule to run its tests - I've pushed a fix for that in 8b49140e5cb6675c58bd908dbd4c7e12146a4f72).
With the apostrophe definition fix CI passes for everything except csharp, which is currently broken in master too. We have other wide character Unicode programming languages passing so I don't expect this hides a problem but I'll try to check it locally and merge the branch if it is indeed OK.
What would it take to have this merged?
@stefanvodita Thanks for your interest.
I've fixed CI on master and pushed a fix for a minor nit here (the new modules.txt
entry wasn't in sorted order) which has demonstrated that the C# failure wasn't anything to do with these changes. This PR looks good now but needs the other two to be ready as well - I'll make some comments on those and @-mention you.
I'm going to close this in favour of #178 as that's been actively worked on.
Looks like you pushed this branch before the corresponding snowball-data branch (the instructions do say to push snowball-data first) and the CI failed because the new testdata wasn't there yet.
I'll trigger it to rerun.