NSSpellChecker certain language not working - objective-c

Question

My mac version is 10.13.4 with xcode version 9.4.1
I'm trying to use NSSpellChecker with Korean language. But regardless of whether the word I check is misspelled or not, the Range returned is {2^63 - 1, 0}
Based on quick experiments with english, that Range is returned when there are no misspelled words OR if the input word language doesn't match current spell checker's language.
This is what I'm doing:
NSSpellChecker *checker = [NSSpellChecker sharedSpellChecker];
NSString *sampleWord = #"마직막";
NSRange range = [checker checkSpellingOfString:sampleWord, startingAt: 0];
NSLog(#"range: %#", NSStringFromRange(range));
that sampleWord is misspelled, but I'm getting
range: {9223372036854775807, 0}
as the output.
Could someone provide any insight on what's happening?
Edit1: When I throw correct english words at it, NSSpellChecker indicates spelling error, which is correct. I gave an english word to korean environment. So I guess NSSpellChecker itself is working correctly, but there is something I missed regarding setup kind of..?

Answer 1

Here is code I used, based on your code and comments.
// insert code here...
NSSpellChecker *checker = [NSSpellChecker sharedSpellChecker];
// NSLog ( #"Available %#", checker.availableLanguages );
NSLog ( #"Korean %#", [checker.availableLanguages containsObject:#"ko"] ? #"YES" : #"NO" ); // Prints YES
[checker setAutomaticallyIdentifiesLanguages:NO]; // Seems to have NO impact
NSLog ( #"KO %#", [checker setLanguage:#"ko"] ? #"YES" : #"NO" ); // Prints YES
// NSString * sampleWord = #"마직막 마직막 마직막 마직막 마직막"; // makes no difference
NSString * sampleWord = #"마직막"; // makes no difference
// NSString * sampleWord = #"마직막 ddd sss aaa"; // consistently prints { 4, 3 }
// NSString * sampleWord = #"안녕하세요 햇살 가득한 남아프리카에서 왔습니다. 마직막 안녕하세요 햇살 가득한 남아프리카에서 왔습니다."; // all fine for this one also
// NSString * sampleWord = #"안녕하세요 햇살 가녕하세득한 남녕하세아프리카에서 왔습니다. 마직막 안녕하세요 햇녕하세살 가득한 남아프리카에서 왔습니다."; // here I finally start to see some result { 45, 5 }
NSRange range = [checker checkSpellingOfString:sampleWord
startingAt:0];
NSLog(#"range: %#", NSStringFromRange(range)); // Consistently prints { NSNotFound, 0 }
NSInteger wc;
// Since you force the language I think this
// is better way of checking
range = [checker checkSpellingOfString:sampleWord
startingAt:0
language:#"ko"
wrap:NO
inSpellDocumentWithTag:0
wordCount:& wc];
NSLog(#"spelling range: %# wc %lu", NSStringFromRange(range), wc); // Consistently prints { NSNotFound, 0 } wc 1
range = [checker checkGrammarOfString:sampleWord
startingAt:0
language:#"ko"
wrap:NO
inSpellDocumentWithTag:0
details:nil];
NSLog(#"grammer range: %# wc %lu", NSStringFromRange(range), wc); // Consistently prints { NSNotFound, 0 } wc 1
The word you are testing with 마직막 appears to be valid and passes consistently - even in the grammar test.
If I create some kind of a sentence and mess it up by randomly pasting in letters then I start getting some sensible (and consistent) feedback from the spell checker. See the code for more info.
Mostly I think you should use, as shown in the code, the message that allows you to specify the language rather than the one that does not since you are forcing the language. But I do not see inconsistency here. If I paste the word into Notes it also does not highlight it, although my intentionally erroneous string does light up there as well.
EDIT
I've played with your code a bit. Note that readline will append the newline so I changed the code by trimming all white space and newlines from theWord as below.
NSString *sampleWord = [[NSString stringWithCString:theWord encoding:NSUTF8StringEncoding] stringByTrimmingCharactersInSet:NSCharacterSet.whitespaceAndNewlineCharacterSet];
Initially I got the discrepancy you mention but just as I got excited it disappeared. So yes, it seems inconsistent. I am not sure if the spell checker got smarter or if the console got better at dealing with the input.
The basic problem, I believe, is that getline reads chars. So I also tried e.g.
NSString *sampleWord = [[[NSString alloc] initWithBytesNoCopy:theWord
length:len_read
encoding:NSUTF8StringEncoding
freeWhenDone:YES] stringByTrimmingCharactersInSet:NSCharacterSet.whitespaceAndNewlineCharacterSet];
but got the same result. Here I also tried
NSASCIIStringEncoding
and
NSUTF16StringEncoding
but only
NSUTF8StringEncoding
worked correctly, so it seems the console returns UTF8 strings. At least for me, I suspect this will be different based on the locale and other settings.
Discrepancy
So what about that. Yesterday when I ran my code, for the very first time, I think it produced a spelling error. Never again. I ran it a lot but never again and I am not even sure it really did produce that error anymore. Same today. I ran your code and the first time it marked the word as an error, but never again. By now, after running it again many times, I am starting to wonder if I was not mistaken as I simply can not generate it again.
Maybe the spell checker takes a while to initialise and this gives rise to the discrepancy. I could not find anything in the manual that points to this. Maybe the console gets smart as I enter Hangul and this is why it generates two separate outputs. Maybe the clipboard sees I am copy and pasting Hangul and changes something somewhere.
The strange thing though is that I repeated the whole thing on my MacBook as well, hoping to get the failure / misspelt word the first time again and this time copying the log out. But there it just worked from the first time.
Maybe run your code a few times and see if it stabilises after a while. Here it seems that (maybe) the first time it marks the word as misspelt but never again, so I am as perplexed as you are at present.
Follow up
Note that I am unable to get the spellchecker to mark the word as misspelt. I also tried the other words you gave and later even some gibberish but the spell checker never flagged anything as incorrect.
Then I tried on my Mac again and this time everything I tried was marked (correctly) as incorrect. So the result is not consistent between my Mac and Macbook, although, where it works, it seems that "마직막" is spelt correctly according to Apple.
I can not explain this but suspect it has to do with language and locale setting differences between the two machines, even though the code is the same and there are no differences I am aware of. On both machines the output is the same, notably it outputs YES when I initialise the Korean dictionary.

Answer 2

This is the code I'm using right now, where "마직막" is failing.
while(true) {
char *theWord = NULL;
size_t len = 0;
printf("ENTER A WORD\n");
size_t len_read = getline(&theWord, &len, stdin);
printf("%s\n", theWord);
NSString *sampleWord = [NSString stringWithCString:theWord encoding:NSUTF8StringEncoding];
NSInteger wc;
NSRange range = [checker checkSpellingOfString:sampleWord startingAt:0 language:#"ko" wrap:NO inSpellDocumentWithTag:0 wordCount:&wc];
if (range.location == NSNotFound) {
printf("Correct Spelling\n");
} else {
printf("Bad Spelling. Incorrect Spelling at\n");
NSLog(#"The range: %#", NSStringFromRange(range));
}
fflush(stdin);
free(theWord);
}
And if I input "마직막", this is what I'm getting.
ENTER A WORD
마직막
ㅁㅏㅈㅣㄱㅁㅏㄱ
Bad Spelling. Incorrect Spelling at
2020-11-27 10:39:23.916315+0900 spellCheckTest[6410:124279] The range: {0, 8}
So we're getting different results using the same input. Could this be because of how I'm accepting input from console??