ReportDataAggregator#readTextFrom seems VERY inefficient

[transentia]

I've been trawling through the code (trying to get a handle on https://github.com/AOEpeople/geb-spock-reports/issues/34).

I came across com.athaydes.spockframework.report.internal.ReportDataAggregator#readTextFrom:

    @PackageScope
    static String readTextFrom( RandomAccessFile file ) {
        def buffer = new byte[8]
        def result = new StringBuilder( file.length() as int )

        int bytesRead
        while ( ( bytesRead = file.read( buffer ) ) > 0 ) {
            result.append( new String( buffer[ 0..( bytesRead - 1 ) ] as byte[], charset ) )
        }
        return result.toString()
    }

Reading 8 bytes at a time and making and appending lots of tiny strings seems really inefficient! Is there perhaps a micro-optimisation that can be done here? Something LIKE:

    static String readTextFrom( RandomAccessFile file ) {
        def buffer = new byte[file.length() as int]

        int bytesRead = file.read(buffer);
        if (bytesRead != buffer.length) {
            // error message or something...
        }

        return new String(buffer, 0, bytesRead, charset)
    }

...this specific code would fail if given a long-sized file, but that's unlikely (? famous last words)?

[renatoathaydes]

I agree, this buffer is too small, and the allocation of a String inside the loop is unnecessary. Do you want to send a PR fixing that?