Fix xls ministream handling.

source/Sylvan.Data.Excel/Packaging/Ole2Package+Ole2Entry.cs

@@ -1,9 +1,7 @@
-using System;
-using System.Collections.Generic;
+using System.Collections.Generic;
 using System.IO;
 using System.Linq;
 using System.Text;
-using System.Text.RegularExpressions;
 
 namespace Sylvan.Data.Excel;
 
@@ -107,8 +105,7 @@ public Stream Open()
  if (this.StreamSize < Package.miniSectorCutoff)
  {
  var sectors = this.Package.GetMiniStreamSectors(this.StartSector).ToArray();
- throw new NotImplementedException();
- //return new Ole2MiniStream(this.Package, sectors, StreamSize);
+ return new Ole2MiniStream(this.Package, this.Package.miniStream, sectors, StreamSize);
  }
  else
  {

source/Sylvan.Data.Excel/Packaging/Ole2Package+Ole2MiniStream.cs

@@ -1,7 +1,6 @@
 using System;
+using System.Diagnostics;
 using System.IO;
-using System.Threading;
-using System.Threading.Tasks;
 
 namespace Sylvan.Data.Excel;
 
@@ -22,9 +21,11 @@ public sealed class Ole2MiniStream : Stream
  uint sector;
 
  long streamPos;
+ Stream miniStream;
 
- public Ole2MiniStream(Ole2Package package, uint[] sectors, long length)
+ public Ole2MiniStream(Ole2Package package, Stream miniStream, uint[] sectors, long length)
  {
+ this.miniStream = miniStream;
  this.package = package;
  this.sectors = sectors;
  this.sectorIdx = 0;
@@ -63,41 +64,40 @@ public override long Seek(long offset, SeekOrigin origin)
  pos = this.length + offset;
  break;
  }
- if (pos < 0 || pos > this.length)
+ if (pos < 0)
  {
  throw new ArgumentOutOfRangeException(nameof(offset));
  }
 
  this.position = pos;
- var idx = pos / this.sectorLen;
+ var idx = pos / MiniSectorSize;
 
  this.sectorIdx = (int)idx;
- this.sectorOff = (int) (pos - (idx * sectorLen));
+ this.sectorOff = (int) (pos - (idx * MiniSectorSize));
  this.sector = this.sectors[sectorIdx];
  return this.position;
  }
 
- public override async Task<int> ReadAsync(byte[] buffer, int offset, int count, CancellationToken cancellationToken)
+ public override int Read(byte[] buffer, int offset, int count)
  {
  if (offset + count > buffer.Length)
  throw new ArgumentOutOfRangeException();
 
- //Debug.WriteLine($"{offset} {count} {this.position}");
-
  var sectors = this.sectors;
 
  int bytesRead = 0;
  var c = count;
+ var z = (int) Math.Min(count, this.length - position);
 
  while (bytesRead < count && position < length)
  {
  var readLen = 0;
- var readStart = (sector + 1) * sectorLen + sectorOff;
+ var readStart = sector * MiniSectorSize + sectorOff;
  var curSector = sector;
 
- while (readLen < c)
+ while (readLen < z)
  {
- if (this.sectorOff >= this.sectorLen)
+ if (this.sectorOff >= MiniSectorSize)
  {
  sectorOff = 0;
  sectorIdx++;
@@ -119,8 +119,9 @@ public override async Task<int> ReadAsync(byte[] buffer, int offset, int count,
  sector = curSector = nextSector;
  }
 
- var sectorAvail = this.sectorLen - this.sectorOff;
- var sectorRead = Math.Min(sectorAvail, c - readLen);
+ var sectorAvail = MiniSectorSize - this.sectorOff;
+ Debug.WriteLine("SA: " + sectorAvail);
+ var sectorRead = Math.Min(sectorAvail, z - readLen);
 
  readLen += sectorRead;
  this.sectorOff += sectorRead;
@@ -129,7 +130,7 @@ public override async Task<int> ReadAsync(byte[] buffer, int offset, int count,
  // avoid seek if we are already positioned.
  if (streamPos != readStart)
  {
- package.stream.Seek(readStart, SeekOrigin.Begin);
+ package.miniStream.Seek(readStart, SeekOrigin.Begin);
  streamPos = readStart;
  }
 
@@ -138,7 +139,7 @@ public override async Task<int> ReadAsync(byte[] buffer, int offset, int count,
  int len = 0;
  while (len < readLen)
  {
- int l = await package.stream.ReadAsync(buffer, offset, readLen).ConfigureAwait(false);
+ int l = package.miniStream.Read(buffer, offset, readLen);
  if (l == 0)
  throw new IOException();//"Unexpectedly encountered end of Ole2Package Stream"
  len += l;
@@ -152,11 +153,6 @@ public override async Task<int> ReadAsync(byte[] buffer, int offset, int count,
  return bytesRead;
  }
 
- public override int Read(byte[] buffer, int offset, int count)
- {
- return ReadAsync(buffer, offset, count, default).GetAwaiter().GetResult();
- }
-
  public override void Flush()
  {
  }

source/Sylvan.Data.Excel/Packaging/Ole2Package+Ole2Stream.cs

@@ -1,7 +1,8 @@
 using System;
+using System.Diagnostics;
 using System.IO;
-using System.Threading;
 using System.Threading.Tasks;
+using System.Threading;
 
 namespace Sylvan.Data.Excel;
 
@@ -44,7 +45,7 @@ public override long Position
  }
  set
  {
- Seek(value - this.position, SeekOrigin.Current);
+ Seek(value, SeekOrigin.Begin);
  }
  }
 

source/Sylvan.Data.Excel/Packaging/Ole2Package.cs

@@ -19,9 +19,11 @@ sealed partial class Ole2Package
  const uint MiniSectorCutoff = 0x1000;
  const int HeaderFatSectorListCount = 109;
  const int DirectoryEntrySize = 0x80;
+ const int MiniSectorSize = 0x40;
 
  BinaryReader reader;
  Stream stream;
+ Stream miniStream; 
 
  int sectorSize;
  ushort verMinor;
@@ -61,19 +63,28 @@ public Ole2Entry GetEntry(int entryIdx)
 
  public Ole2Package(Stream iStream)
  {
-
  this.stream = iStream;
  this.reader = new BinaryReader(iStream, Encoding.Unicode);
  this.fatSectorList = Array.Empty<uint>();
  this.miniFatSectorList = Array.Empty<uint>();
  this.entryList = Array.Empty<Ole2Entry>();
  LoadHeader();
  LoadDirectoryEntries();
+
+ if (RootEntry.StartSector <= MaxSector)
+ {
+ var miniStreamSectors = GetStreamSectors(RootEntry.StartSector).ToArray();
+
+ this.miniStream = new Ole2Stream(this, miniStreamSectors, RootEntry.StreamSize);
+ }
+ else
+ {
+ this.miniStream = Stream.Null;
+ }
  }
 
  void LoadHeader()
  {
-
  BinaryReader reader = new BinaryReader(stream, Encoding.Unicode);
 
  ulong magic = reader.ReadUInt64();
@@ -123,7 +134,7 @@ void LoadHeader()
 
  this.fatSectorCount = reader.ReadUInt32();
 
- directorySectorStart = reader.ReadUInt32();
+ this.directorySectorStart = reader.ReadUInt32();
 
 
  uint sig = reader.ReadUInt32();
@@ -209,7 +220,7 @@ void LoadMiniFatSectorList()
  int diCount = sectorSize / 4 - 1;
 
  for (int k = 0; k < diCount; k++)
- fatSectorList[i++] = reader.ReadUInt32();
+ miniFatSectorList[i++] = reader.ReadUInt32();
 
  // read next difat location
  sect = reader.ReadUInt32();
@@ -247,32 +258,12 @@ IEnumerable<uint> GetMiniStreamSectors(uint startSector)
  do
  {
  yield return sector;
- sector = NextMiniSector(sector);
+ sector = miniFatSectorList[sector];
  if (sector == startSector || sector == 0)
  throw new InvalidDataException();
  } while (sector != EndOfChain);
  }
 
- public uint NextMiniSector(uint sector)
- {
- throw new NotImplementedException();
- //uint fatSectIdx = (uint)(sector / (sectorSize / 4));
- //uint fatSectOff = (uint)(sector % (sectorSize / 4));
-
- //uint fatPage = miniSectorCutoff[fatSectIdx];
-
- //stream.Seek(SectorOffset(fatPage) + (fatSectOff * 4), SeekOrigin.Begin);
- //return reader.ReadUInt32();
- }
-
-
- IEnumerable<Ole2Entry> GetEntries()
- {
- var root = this.RootEntry;
- foreach (var entry in EnumerateEntry(root))
- yield return entry;
- }
-
  IEnumerable<Ole2Entry> EnumerateEntry(Ole2Entry entry)
  {
  yield return entry;

source/Sylvan.Data.Excel/Xls/XlsWorkbookReader+RecordReader.cs

@@ -10,7 +10,7 @@ sealed partial class XlsWorkbookReader
 {
  sealed class RecordReader
  {
- const int BufferSize = 0x40000;
+ const int BufferSize = 0x10000;
  const int MaxRecordSize = 8228;
 
  Stream stream;