Re: AW: httpapi speed

[Scott Klement <sk@xxxxxxxxxxxxxxxx>]

Yes, I agree that your modification should do no harm.   It would either 
improve performance or have no impact on performance.  In fact, the 
original code in HTTPAPI worked like that (before the additional header 
block grew so large.)

So I will make your modification to HTTPAPI.

I just wanted to know how disabling Nagle's algorithm worked for you, 
since that should improve all requests.  I was pretty sure that most of 
the delay was coming from Nagle's algorithm, and wanted to confirm it.


On 5/31/2014 3:45 PM, Sawatzki, Peter wrote:
> Notice my modification should do no harm: if additional headers are present, it would send the headers and the CRLF at the end in separate requests. However if no addtl headers are present, the request and the terminating CRLF are sent in one paket (usually).

-----------------------------------------------------------------------
This is the FTPAPI mailing list.  To unsubscribe, please go to:
http://www.scottklement.com/mailman/listinfo/ftpapi
-----------------------------------------------------------------------