Regex behavior in Falko metadata is incorrect
Bug #967096 reported by
Amir Zeldes
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
ANNIS |
Fix Committed
|
Undecided
|
Thomas Krause |
Bug Description
When searching for regex in annotations, it is assumed the user means an initially and finally anchored string, i.e. the query
tok=/[Uu]nd/
is actually looking for the expression /^[Uu]nd$/. As a consequence, this query does not find "Hund". This is correct.
In metadata, this does not happen correctly. Looking in FalkoEssaysL2, this query:
tok & meta::l1_1=/[^e].+/
finds texts with l1_1="eng", because "eng" contains [^e].+ as a substring: the "ng" part. This query should also be treated as looking for l1_1=/^[^e].+$/, so that metadata and annotations behave in the same way.
Related branches
Changed in annis: | |
milestone: | none → 3.0.0 |
status: | New → In Progress |
Changed in annis: | |
status: | In Progress → Fix Committed |
To post a comment you must log in.