Small problems with tcpmon

[co...@covalent.net]

Hi,

I run into 2 small bugs - the diff solves the problem for me,
maybe for others :-)

First, on some VMs the SocketRR will hung for a very long time.
That's because even if close() is called on the socket, the
read() will remain blocked. Adding a soTimeout and 
checking for done solves this.

The other - with some servers, if available() returns 
0 we read a single byte, and then send it. Then we'll
send a second packet with the rest of the request.
That will confuse at least apache  (who would see
OST instead of POST - still trying to figure out why ). 
I see no reason to read a single byte - we'll block
anyway. 

Costin

Re: Small problems with tcpmon

[co...@covalent.net]

Any problem with my patch ? What's the right procedure, should I file
a bug ?

Costin

On Tue, 26 Feb 2002 costinm@covalent.net wrote:

> I run into 2 small bugs - the diff solves the problem for me,
> maybe for others :-)
> 
> First, on some VMs the SocketRR will hung for a very long time.
> That's because even if close() is called on the socket, the
> read() will remain blocked. Adding a soTimeout and 
> checking for done solves this.
> 
> The other - with some servers, if available() returns 
> 0 we read a single byte, and then send it. Then we'll
> send a second packet with the rest of the request.
> That will confuse at least apache  (who would see
> OST instead of POST - still trying to figure out why ). 
> I see no reason to read a single byte - we'll block
> anyway. 
> 
> Costin
> 
> 
> 
> 
>