Thanks to contributions from Bill Garrison, I’ve updated BSJSONAdditions to 1.2. Bill adapted the unit tests from Jonathan Wight’s CocoaJSON and found that my code was not correctly parsing escaped Unicode characters. So Bill provided a fix along with the adapted unit tests. The updated code is available from the newly-created, overly-colorful Downloads page.
Under what license are you releasing this? I see no mention in any of your files.
Thanks!
Some of the older code isn’t tagged with a license, but consider it an MIT-style license.
Great library Blake. I found this to be incredibly useful. Just wanted to drop by and say thanks.
Hey Blake, great software! Found a bug in it though, which I emailed you about, but in case the email doesn’t get to you I thought I’d post it here so people can fix it themselves if they so wish.
On line 160 of NSScanner BSJSONAdditions.m, the hex character set only includes numbers and uppercase letters– if you’re having problems with Unicode characters not being properly recognized, just add the lowercase “abcdef” to the end of the character set. Problem solved.
Thanks again for some awesome software.
Found a bug in it though. This showed up in our unit tests, when passing the empty string:
Posting the update here:
In:
- (BOOL)scanJSONValue:(id *)value
{
BOOL result = NO;
[self scanJSONWhiteSpace];
add:
if ([self isAtEnd]) {
return result;
}