English spellchecker should not attempt to check words in non-English alphabets.



Message Compose Window
14 years ago
14 years ago


(Reporter: Adam Rice, Assigned: Scott MacGregor)


Firefox Tracking Flags

(Not tracked)



(1 attachment)



14 years ago
User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040614 Moonbird/0.8 (All your Firefox/0.8 are belong to Firesomething)
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040614 Moonbird/0.8 (All your Firefox/0.8 are belong to Firesomething)

My job entails email communication in both English and Japanese. I work with
Thunderbird set to always spellcheck, and it is a constant help. But when I send
a message in Japanese, the spellchecker attempts to tokenise and spellcheck the
email as if it was English! This looks ridiculous. Shouldn't the tokeniser know
what characters it can handle and ignore ones that it can't?

Reproducible: Always
Steps to Reproduce:
1.Enable "Check spelling before sending"
2.Write an email in Japanese (or other non-latin alphabet language)
3.Click "Send"

Actual Results:  
The spellchecker claimed the first chunk of my email was incorrectly spelled,
and offered to add it to my English-language dictionary.

Expected Results:  
Ignored the characters it didn't understand.

Comment 1

14 years ago
Created attachment 155908 [details]
Screenshot illustrating the problem

Comment 2

14 years ago
This problem has been replaced by another one (Bug 280512). I'm closing it as
WORKSFORME. Everyone on the CC list are welcomed to CC themselves to the other
bug - I promise it's just as annnoying and effects the same audience ;-)


Last Resolved: 14 years ago
Resolution: --- → WORKSFORME

Comment 3

14 years ago
This bug is solvable when it comes to texts in languages with disjoint character
sets, such as English and Japanese. However, if you have a French paragraph
(without accents) in an English text, it becomes impossible to skip French words
according to the character set.

I believe that a good solution would be to let the user select a *set* of
spelling languages - not just one as is done now. Then, if a word is valid in
one of the languages selected, it is considered correct.

This would be helpful, for example, for Hebrew technical writers, who tend to
use English terms. My joint Hebrew+English dictionary
http://ivrix.org.il/redhat/heb-eng.xpi was made for that reason, and I wait for
the day it becomes redundant.

Comment 4

14 years ago
as always happens in bugzilla, my lament for multilingual speller appears
already in bug 227610, and mentioned on bug 69687.
You need to log in before you can comment on or make changes to this bug.