Back to LanguageTool Homepage - Privacy - Imprint

French: false positive report of missing non-breaking space


(Didier Briel) #1

Hello,

I use LanguageTool 3.2 embedded in OmegaT development version.

Whenever I enter a "double" punctuation sign, I have the warning "Point d'interrogation (or whatever sign it is) est précédé d'une espace fine insécable" (Question mark should have a leading small non-breaking space) even if I actually use a non-breaking space.

Previously, OmegaT used LanguageTool 2.2 and it works fine.

If I try the online test page I have the same false positive, so it seems unlikely the error comes from OmegaT embedding.

I thought perhaps the rule had become more strict, really requiring a small ("fine") non-breaking space, but that doesn't seem the case according to the source code:
github.com/languagetool-org/languagetool/blob/master/languagetool-language-modules/fr/src/main/java/org/languagetool/rules/fr/QuestionWhitespaceRule.java

// Strictly speaking, the character before ?!;: should be an
// "espace fine insécable" (U+202f). In practise, an
// "espace insécable" (U+00a0) is also often used. Let's accept both.

Does someone else has the issue?

Thank you for your attention,

Didier


(Aaron Madlon-Kay) #2

This appears to be either a bug with the rule or bad guidance in the documentation regarding the required Unicode normalization of the input text: