The improvement was very dramatic with moving to your code. I think I see
why my code was so slow, I completely forgot about the boxing overhead. It
makes quite a bit of difference.
Erick
----- Original Message -----
From: "Erick Thompson" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, May 03, 2002 5:24 PM
Subject: Re: [DOTNET] Downloading a binary file using HttpWebResponse
> Sean,
>
> Thanks for the code, this looks like it should be much faster. I may
create
> a strongly typed byte ArrayList, which would reduce the memory copies from
> O(n) to O(log n), if I get enough time.
>
> Erick
>
> ----- Original Message -----
> From: "Sean Greer (SBI-Chico)" <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>
> Sent: Friday, May 03, 2002 4:13 PM
> Subject: Re: [DOTNET] Downloading a binary file using HttpWebResponse
>
>
> > Here's my stab at this problem...this is off the top of my head and has
> not
> > been thoroughly tested, but it does work against my pseudo web service.
> :-)
> > YMMV
> >
> > byte[] arrBuffer = new byte[0];
> > using(BinaryReader reader = new
> > BinaryReader(webResponse.GetResponseStream()))
> > {
> > byte[] arrScratch = null;
> > while((arrScratch = reader.ReadBytes(4096)).Length > 0)
> > {
> > if(arrBuffer.Length == 0)
> > arrBuffer = arrScratch;
> > else
> > {
> > byte[] arrTemp = new byte[arrBuffer.Length +
> > arrScratch.Length];
> > Array.Copy(arrBuffer, arrTemp,
arrBuffer.Length);
> > Array.Copy(arrScratch, 0, arrTemp,
> arrBuffer.Length,
> > arrScratch.Length);
> > arrBuffer = arrTemp;
> > }
> > }
> > }
> > // use arrBuffer now...
> >
> > HTH,
> >
> > Seang
> >
> > -----Original Message-----
> > From: Erick Thompson [mailto:[EMAIL PROTECTED]]
> > Sent: Friday, May 03, 2002 2:24 PM
> > To: [EMAIL PROTECTED]
> > Subject: [DOTNET] Downloading a binary file using HttpWebResponse
> >
> >
> > I'm downloading binary files from the Stream returned by a
HttpWebResponse
> > object. The stream returned does not support seeking or look aheads. I
> have
> > no idea when the length of the stream will be
> (HttpWebResponse.ContentLength
> > cannot be trusted). What is the fastest way to get the stream into an
> array
> > of bytes?
> >
> > Below is the code that I'm using right now. It does work, but it's slow
> and
> > memory intensive (due to the ArrayList and the cast to a byte array).
> There
> > must be a better way.
> >
> > Thanks,
> > Erick
> >
> > private void LoadBinaryContent(HttpWebResponse res) {
> > Stream binStream = res.GetResponseStream();
> > BinaryReader reader = new BinaryReader(binStream);
> > int len = (int)res.ContentLength;
> > if (len == 0) {
> > len = 100000; // make a guess
> > }
> > ArrayList tmp = new ArrayList(len);
> > try {
> > while (true) {
> > tmp.Add(reader.ReadByte());
> > }
> > } catch (EndOfStreamException) {
> > byte b = 0;
> > _binaryContent = (byte[])tmp.ToArray(b.GetType());
> > }
> > reader.Close();
> > binStream.Close();
> > }
> >
> > You can read messages from the DOTNET archive, unsubscribe from DOTNET,
or
> > subscribe to other DevelopMentor lists at http://discuss.develop.com.
> >
> > You can read messages from the DOTNET archive, unsubscribe from DOTNET,
or
> > subscribe to other DevelopMentor lists at http://discuss.develop.com.
>
> You can read messages from the DOTNET archive, unsubscribe from DOTNET, or
> subscribe to other DevelopMentor lists at http://discuss.develop.com.
You can read messages from the DOTNET archive, unsubscribe from DOTNET, or
subscribe to other DevelopMentor lists at http://discuss.develop.com.
