New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Locale parsing issues #13229
Comments
@javanna how (and how often) does it break bwc? |
Some locales that used to get properly parsed might not get parsed correctly anymore. I am not sure how widely these options are used, especially in queries. Maybe removing this option as stated in #9978 would be the best solution. |
Yes, using the user's already configured analysis chain rather than applying additional configuration (lowercase_expanded_terms, locale, etc) that is still limited to only case is really a better way to go IMO. If we solve that issue then if they have turkishlowercase, it gets used, if they use asciifolding, it gets used, and so on, and people stop getting confused about why wildcards "don't work" which I bet happens constantly. |
seems like the way forward would be to remove these |
…ry_string The configured analysis chain should be used instead. Relates to elastic#13229
while looking into removing |
Closing in favour of #9978. Once we have made it possible to use the analysis chain, by taking out components that are not multi term aware, we can go ahead and remove |
In the query-refactoring branch we noticed weird behaviours related to parsing locales. The two queries affected are query_string and simple_query_string, as they allow users to set the
locale
used where applicable, depending on the type of lucene query that gets parsed.Both queries in master parse the locale through our own
LocaleUtils#parse
method and print it out usingLocale#toString
when needed. It turns out that with some locales, our newly added query tests fail given that printing out and reparsing the same locale leads to a different locale. The problem looks very similar to what is explained here: https://issues.apache.org/jira/browse/LUCENE-4021.I see that we also use
LocaleUtils#parse
inDateFieldMapper
. I think that we should discuss how to address this issue and streamline the solution. I think the best solution would be to parse throughLocale#forLanguageTag
and print out usingLocale#toLanguageTag
, but this breaks backwards compatibility it seems. If I make this changeSimpleDateMappingTests#testParseLocal
fails for instance. Is also our ownLocaleUtils
still useful?The text was updated successfully, but these errors were encountered: