Closed robcowie closed 12 years ago
Right now, input (matched messages) is assumed to be utf8
and decoded with the ignore
flag.
No attempt is made to detect the encoding of the target file. This is obviously not ideal behaviour but it does at least prevent failures due to encoding issues. I'll revisit this stuff it it proves to be an issue.
When building the result output, the matched string is implicitly decoded. It fails if it cannot be decoded as ascii.
Is there a fast way of determining the encoding of a file? If not, perhaps just try utf8, else use ST2
fallback_encoding