System.OutOfMemoryException when tried with 70 files

[surensaluka]

Steps to reproduce

1. Use 70 different pdf files which are in various sizes. (Combined size is 165MB)
2. Try looping through all 70 files adding to ZipOutputStream
3. Throw OutOfMemoryException
4. Use my sample code

public byte[] DownloadPaperList()
{
    try
    {
        PaperDAO paperDAO = new PaperDAO();
        List<PaperModel> papers = paperDAO.GetPaperListByPaperIds(paperIDs);
        using (MemoryStream outputMemoryStream = new MemoryStream())
        {
            using (var zipOutputStream = new ZipOutputStream(outputMemoryStream))
            {
                papers = papers.OrderBy(x => x.Order).ToList();
                foreach (PaperModel paper in papers)
                {
                    byte[] decryptedPaper = CryptoServices.DecryptFile(
                        paper.FileData, 
                        paperDAO.GenerateCloseOpenFile(paper, string.Empty), 
                        paper.SEVersion);

                    ZipEntry zipFileEntry = new ZipEntry(paper.DocName + ".pdf")
                    {
                        Size = decryptedPaper.Length
                    };

                    zipOutputStream.SetLevel(3);

                    //EXCEPTION THROWS HERE!!!
                    zipOutputStream.PutNextEntry(zipFileEntry); 

                    StreamUtils.Copy(new MemoryStream(decryptedPaper), 
                        zipOutputStream, 
                        new byte[4096]);
                }
                zipOutputStream.CloseEntry();
                // Stop ZipStream.Dispose() from also Closing the underlying stream.
                zipOutputStream.IsStreamOwner = false;
                outputMemoryStream.Position = 0;
            }
            return outputMemoryStream.ToArray();
        }
    }            
    catch (Exception)
    {
        throw;
    }
}

Expected behavior

Create a zip file and return the result as a byte[]

Actual behavior

Throw OutOfMemoryException

Version of SharpZipLib

ICSharpCode.SharpZipLib 0.85.4.369

Obtained from (only keep the relevant lines)

• Package installed using NuGet

[piksel]

This is probably due to random allocation of additional memory. Try setting the initial size of the outputMemoryStream to something close to what the final size ought to be.
It could also be because of how the decompression is attempted inside SharpZipLib, in that case I would suggest trying "Store" instead of Deflation compression, as PDFs usually doesn't compress that well anyway (really depends on the content though).
Another thing you could try is to write the files to a temporary folder before compressing them. If the compression is done in a separate loop, chances are that the memory allocated for the decryption could all be freed by the GC, allowing the compression to allocate the necessary memory chunks.

[Numpsy]

Probably pretty trivial, but maybe try writing the decryptedPaper byte array straight into the output stream instead of creating temporary memory streams and buffers inside the loop?

[piksel]

Closing due to lack of response.