Matching 8-bit characters in Haskell Text.Regex.Posix.ByteString

Question

my Haskell application reads input as a list of ByteString and I'm using Text.Regex.Posix.ByteString.regexec to find matches. Some input has a character code 253 (it's a 1/2 symbol in one IBM PC character set) and it seems that the pattern '.' (i.e., dot, "match any character") doesn't match it. Any way to make it match ?

score 0 · Answer 1 · answered Oct 25 '13 at 07:29

0

That doesn't make sense. Why would you want to match a half-character? . will match the full character.

answered Oct 25 '13 at 07:29

David Knipe

3,417
1
19
19

It's only half a character in certain encodings such as UTF-8. I'm looking to match 8-bit characters. – Michael restore Monica Cellio Oct 25 '13 at 07:31

score 0 · Accepted Answer · answered Mar 26 '15 at 16:08

This works for me on a Windows Haskell install:

> length $ ((pack ['\1'..'\253']) =~ "." :: [[ByteString]])
252

I.e. dot matches all characters in range including code 253.

Note that the library calls out to the underlying posix regex matcher, typically, I assume, from glibc.

So I would imagine any issue you have would be with that precise underlying c implementation.

Something like Text.Regex.TDFA.ByteString might give you clearer behavior in this case, since it is all in Haskell?

Matching 8-bit characters in Haskell Text.Regex.Posix.ByteString

2 Answers2