/// <summary>Seek to the first line that starts immediately before 'filepos'</summary>
private static long FindLineStart(IFileSource file, long filepos, long fileend, byte[] row_delim, Encoding encoding, byte[] buf)
{
file.Stream.Seek(filepos, SeekOrigin.Begin);
// Read a block into 'buf'
bool eof;
int read = Buffer(file, buf.Length, fileend, encoding, true, buf, out eof);
if (read == 0)
{
return(0); // assume the first character in the file is the start of a line
}
// Scan for a line start
int idx = Misc.FindNextDelim(buf, read - 1, read, row_delim, true);
if (idx != -1)
{
return(file.Stream.Position + idx); // found
}
if (filepos == read)
{
return(0); // assume the first character in the file is the start of a line
}
throw new NoLinesException(read);
}
/// <summary>Populate this line from a buffer</summary>
public void Read(long addr, byte[] buf, int start, int length, Encoding encoding, byte[] col_delim, List <Highlight> highlights, IEnumerable <Transform> transforms)
{
LineStartAddr = addr;
// Convert the buffer to text
RowText = encoding.GetString(buf, start, length);
//RowText = RowText.TrimEnd(new[]{'\r','\n'}); - don't do this, its the logs fault if it has weird newlines at the end of each row
// Apply any transforms
foreach (var tx in transforms)
{
RowText = tx.Txfm(RowText);
}
Column.Clear();
// Split the line into columns
if (col_delim.Length == 0) // Single column
{
Column.Add(new Col(RowText, highlights));
}
else // Multiple columns
{
int e, s = start;
do
{
e = Misc.FindNextDelim(buf, s, length, col_delim, false); // Returns one passed the delimiter
var col_text = encoding.GetString(buf, s, e - s - (e != length?col_delim.Length:0));
Column.Add(new Col(col_text, highlights));
s = e;
}while (e != length);
}
}
/// <summary>Scan the file from 'filepos' adding whole lines to 'line_index' until 'length' bytes have been read or 'add_line' returns false</summary>
/// <param name="file">The file to scan</param>
/// <param name="filepos">The position in the file to start scanning from</param>
/// <param name="fileend">The current known length of the file</param>
/// <param name="backward">The direction to scan</param>
/// <param name="length">The number of bytes to scan over</param>
/// <param name="add_line">Callback function called with each detected line</param>
/// <param name="encoding">The text file encoding</param>
/// <param name="row_delim">The bytes that identify an end of line</param>
/// <param name="buf">A buffer to use when buffering file data</param>
/// <param name="progress">Callback function to report progress and allow the find to abort</param>
private static void FindLines(IFileSource file, long filepos, long fileend, bool backward, long length, AddLineFunc add_line, Encoding encoding, byte[] row_delim, byte[] buf, ProgressFunc progress)
{
long scanned = 0, read_addr = filepos;
for (;;)
{
// Progress update
if (progress != null && !progress(scanned, length))
{
return;
}
// Seek to the start position
file.Stream.Seek(read_addr, SeekOrigin.Begin);
// Buffer the contents of the file in 'buf'.
long remaining = length - scanned; bool eof;
int read = Buffer(file, remaining, fileend, encoding, backward, buf, out eof);
if (read == 0)
{
break;
}
// Set iterator limits.
// 'i' is where to start scanning from
// 'iend' is the end of the range to scan
// 'ilast' is the start of the last line found
// 'base_addr' is the file offset from which buf was read
int i = backward ? read - 1 : 0;
int iend = backward ? -1 : read;
int lasti = backward ? read : 0;
long base_addr = backward ? file.Stream.Position : file.Stream.Position - read;
// If we're searching backwards and 'i' is at the end of a line,
// we don't want to count that as the first found line so adjust 'i'.
// If not however, then 'i' is partway through a line or at the end
// of a file without a row delimiter at the end and we want to include
// this (possibly partial) line.
if (backward && IsRowDelim(buf, read - row_delim.Length, row_delim))
{
i -= row_delim.Length;
}
// Scan the buffer for lines
for (i = Misc.FindNextDelim(buf, i, read, row_delim, backward); i != iend; i = Misc.FindNextDelim(buf, i, read, row_delim, backward))
{
// 'i' points to the start of a line,
// 'lasti' points to the start of the last line we found
// Get the range in buf containing the line
RangeI line = backward
? new RangeI(i, lasti - row_delim.Length)
: new RangeI(lasti, i - row_delim.Length);
// Pass the detected line to the callback
if (!add_line(line, base_addr, fileend, buf, encoding))
{
return;
}
lasti = i;
if (backward)
{
i -= row_delim.Length + 1;
}
}
// From 'lasti' to the end (or start in the backwards case) of the buffer represents
// a (possibly partial) line. If we read a full buffer load last time, then we'll go
// round again trying to read another buffer load, starting from 'lasti'.
if (read == buf.Length)
{
// Make sure we're always making progress
long scan_increment = backward ? (read - lasti) : lasti;
if (scan_increment == 0) // No lines detected in this block
{
throw new NoLinesException(read);
}
scanned += scan_increment;
read_addr = filepos + (backward ? -scanned : +scanned);
}
// Otherwise, we're read to the end (or start) of the file, or to the limit 'length'.
// What's left in the buffer may be a partial line.
else
{
// 'i' points to 'iend',
// 'lasti' points to the start of the last line we found
// Get the range in buf containing the line
RangeI line = backward
? new RangeI(i + 1, lasti - row_delim.Length)
: new RangeI(lasti, i - (IsRowDelim(buf, i - row_delim.Length, row_delim) ? row_delim.Length : 0));
// ReSharper disable RedundantJumpStatement
// Pass the detected line to the callback
if (!add_line(line, base_addr, fileend, buf, encoding))
{
return;
}
// ReSharper restore RedundantJumpStatement
break;
}
}
}
