RE Lexer improperly parsing dashes after character classes

[Sevaarcen]

When using a regular expression containing a character class, followed by a dash -, followed by anything else, Yara fails with the following error:

$ /usr/bin/yara -s ~/testrule.yara ~/yara_test_file.txt 
~/testrule.yara(3): error: invalid regular expression "$": bad character range

It appears this error message is related to the following section of the RE lexer, where it doesn't properly look for the ending square bracket of the character range and therefore over-matches on the dash outside.

yara/libyara/re_lexer.l

Line 311 in 7e00915

<char_class>(\\x{hex_digit}{2}|\\.|[^\\])\-[^]] {

This is the test rule I'm using, and the test file being scanned.

$ cat ~/testrule.yara 
rule test {
        strings:
                $ = /[0-9]-2/
        condition:
                all of them
}

$ cat ~/yara_test_file.txt 
1-2-3-4-5

Using grep, regex101, or similar, we can confirm the regex works as expected and is valid.

$ cat ~/yara_test_file.txt | grep -o "[0-9]-2"
1-2

[plusvic]

Fixed in a91b8a1 and ee96a35.