Hi Denis,
Please open a bug report here:
https://issues.apache.org/jira/browse/CMIS
I'll take a look at it.
Thanks,
Florian
Hi Everyone
There is an issue in the method
DotCMIS.Binding.AtomPub.AtomEntryWriter.WriteContent()
private void WriteContent(XmlWriter writer)
{
using (var br = new BinaryReader(stream))
{
var buffer = new byte[BufferSize];
int readBytes = 0;
do
{
readBytes = br.Read(buffer, 0, BufferSize);
writer.WriteBase64(buffer, 0, readBytes);
} while (BufferSize <= readBytes);
}
}
The problem is that the method "Stream.Read() or BinaryReader.Read()"
may
read less amount of bytes than the provided buffer size when the end
of the
stream has not been reached yet.
For example.
The file "File1.txt" of size 30K is requested by HttpWebRequest to
download
from a server and transfer it to Alfresco.
HttpWebRequest returns System.Net.ConnectStream.
The buffer size to read from stream is 65K.
In the code above the first read operation returns 4K from the stream
and
the condition "while (BufferSize <= readBytes)" finishes the read
cycle.
I've fixed the problem with the code below. The read cycle should
continue
working unit Read returns the value != 0:
private void WriteContent(XmlWriter writer)
{
using (var br = new BinaryReader(stream))
{
var buffer = new byte[BufferSize];
int readBytes = 0;
while ((readBytes = br.Read(buffer, 0, BufferSize))
!= 0)
{
writer.WriteBase64(buffer, 0, readBytes);
}
/*
do
{
readBytes = br.Read(buffer, 0, BufferSize);
writer.WriteBase64(buffer, 0, readBytes);
} while (BufferSize <= readBytes);
*/
}
}
Would you please consider this code and update DotCMIS library.
Best regards,
Denis Andreev
