«

They r two processes, for client and server, connected with TCP. Connection
is OK.
I killed the server process, while the client seems not to feel that before
it sent some data by calling send().
And it checked the errno, which indicated a segmentation fault error.

Why should it be segmentation fault rather than something indicating
disconnection?

On 27 Jun, 11:06, "Jimmy" <lofe...@yahoo.com.cn> wrote:
> They r two processes, for client and server, connected with TCP. Connection
> is OK.
> I killed the server process, while the client seems not to feel that before
> it sent some data by calling send().
> And it checked the errno, which indicated a segmentation fault error.
>
> Why should it be segmentation fault rather than something indicating
> disconnection?

standard C has no support for TCP you need to ask on a platform
specific group (eg. Windows or Unix). It might be an idea to post
your code when you do that. I suspect the client is ignoring some
error indication.


--
Nick Keighley

> standard C has no support for TCP you need to ask on a platform
> specific group (eg. Windows or Unix). It might be an idea to post
> your code when you do that. I suspect the client is ignoring some
> error indication.
>
>
> --
> Nick Keighley
>

Well, I realize the segmentation fault was caused by perror(errno), in which
errno has the value of 29. Is that an invalid number?

In article <f5tmkv$78u$>, Jimmy <> wrote:

>Well, I realize the segmentation fault was caused by perror(errno), in which
>errno has the value of 29. Is that an invalid number?

Error numbers are implementation defined.

A segmentation fault in perror() may well indicate some unrelated
memory corruption, or that you have closed stderr.

-- Richard
--
"Consideration shall be given to the need for as many as 32 characters
in some alphabets" - X3.4, 1963.

On 27 Jun, 12:55, "Jimmy" <lofe...@yahoo.com.cn> wrote:

- leave in attributions
- leave in what you are replying to

> > standard C has no support for TCP you need to ask on a platform
> > specific group (eg. Windows or Unix). It might be an idea to post
> > your code when you do that. I suspect the client is ignoring some
> > error indication.
>
> > --
> >Nick Keighley

- don't quote signatures


> Well, I realize the segmentation fault was caused by perror(errno), in which
> errno has the value of 29. Is that an invalid number?

which part of "you need to ask on a platform specific group" didn't
you understand?

--
Nick Keighley

My god it's full of stars!
Dave Bowman, on seeing HAL's source code

Nick Keighley <> writes:

> On 27 Jun, 12:55, "Jimmy" <lofe...@yahoo.com.cn> wrote:
>
> - leave in attributions
> - leave in what you are replying to
>
>> > standard C has no support for TCP you need to ask on a platform
>> > specific group (eg. Windows or Unix). It might be an idea to post
>> > your code when you do that. I suspect the client is ignoring some
>> > error indication.
>>
>> > --
>> >Nick Keighley
>
> - don't quote signatures
>
>
>> Well, I realize the segmentation fault was caused by perror(errno), in which
>> errno has the value of 29. Is that an invalid number?
>
> which part of "you need to ask on a platform specific group" didn't
> you understand?

Which part of "my god, haven't you become a bossy little twit" do you
not understand?

Richard wrote:
> Nick Keighley <> writes:
>> On 27 Jun, 12:55, "Jimmy" <lofe...@yahoo.com.cn> wrote:
>>
>> - leave in attributions
>> - leave in what you are replying to
>>
>>>> standard C has no support for TCP you need to ask on a platform
>>>> specific group (eg. Windows or Unix). It might be an idea to post
>>>> your code when you do that. I suspect the client is ignoring some
>>>> error indication.
>>>
>>>> --
>>>>Nick Keighley
>>
>> - don't quote signatures
>>
>>> Well, I realize the segmentation fault was caused by perror(errno),
>>> in which errno has the value of 29. Is that an invalid number?
>>
>> which part of "you need to ask on a platform specific group" didn't
>> you understand?
>
> Which part of "my god, haven't you become a bossy little twit" do you
> not understand?

You are obviously totally unaware that contravention of standard
Usenet practice needs to be stamped on (and pointed out) at the
earliest possible moment. "Jimmy" is obviously a new user, and
unaware, and in need of information. Maybe also reading lessons.

--
<http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>
<http://www.securityfocus.com/columnists/423>
<http://www.aaxnet.com/editor/edit043.html>
cbfalconer at maineline dot net



--
Posted via a free Usenet account from http://www.teranews.com

"Jimmy" <> wrote:
# They r two processes, for client and server, connected with TCP. Connection
# is OK.
# I killed the server process, while the client seems not to feel that before
# it sent some data by calling send().
# And it checked the errno, which indicated a segmentation fault error.
#
# Why should it be segmentation fault rather than something indicating
# disconnection?

Most likely problem is your error-handling code is blown. You need
to capture the failure and find out where it is happenning. If you
have a debugger, that should point directly to the fault; otherwise
you can sprinkle printfs through the code to verify it is operating
as expected.

If you are using signal handlers, it's not safe to do I/O from there.

--
SM Ryan http://www.rawbw.com/~wyrmwif/
We found a loophole; they can't keep us out anymore.

>> standard C has no support for TCP you need to ask on a platform
>> specific group (eg. Windows or Unix). It might be an idea to post
>> your code when you do that. I suspect the client is ignoring some
>> error indication.
>>
>>
>> --
>> Nick Keighley
>>
>
>Well, I realize the segmentation fault was caused by perror(errno), in which
>errno has the value of 29. Is that an invalid number?

The first argument of perror() is supposed to be a *STRING*, which
errno is not. This probably caused a segmentation fault. Guess what,
nitpickers, this actually is a question about C.

> The first argument of perror() is supposed to be a *STRING*, which
> errno is not. This probably caused a segmentation fault. Guess what,
> nitpickers, this actually is a question about C.
>

Thanks a lot. I done made a blunt mistake!