I am writing a class that will read lines from a log file when it is updated.
I am using Apache VFS2 to get a method called when a file is updated. My main issue is I don't want to read the line from the file if the line is not complete yet, as in it does have a "\n" or "\r" line separator type character at the end. I think i have looked at all the Java libraries i can to read lines but they all discard the EOF and line termination information so I don't think I can use them.
Instead I am looking at reading it in byte by byte and then checking the result to then discard all stuff that comes after the last line separator. I was wondering what you folks thoughts on the best method for doing this is.
So for example:
2013-Jul-01_14:07:17.875 - Connection to Message Bus is reestablished<LF>
2013-Jul-01_14:07:17.875 - Connection to Message Bus is reestablished<LF>
2013-Jul-01_14:15:08.205 - No connection to Message Bus - reestablish before we can publish<LF>
2013-Jul-01_14:15:08.205 - NOT A REAL LINE PLEASE DONT READ
I want to read in the first 3 but not the fourth as it doesn't have a line feed or carriage return character ().
I have looked at Apache commons-io Tailer stuff but I cant tell if that will give me "incomplete" lines (and I realize I will have to ditch the VFS2 stuff to use it).
So psudo-code:
private void ingestFileObject(FileObject file) {
BufferedInputStream bs = new BufferedInputStream(file.getContent().getInputStream());
StringBuilder result = new StringBuilder();
while (bs.available() > 0) {
result.append((char) bs.read());
}
bs.close();
String resultString = result.toString();
//determine what part of resultString is after last carriage return/line seperate (using regex [\\r\\n]+?
//remove the offending part of String.
}
Or any other solutions completely ignoring my psudo-code are welcome at this point too...
Thanks