From: Rainer Weikusat on
"Ersek, Laszlo" <lacos(a)caesar.elte.hu> writes:
> On Tue, 4 May 2010, boltar2003(a)boltar.world wrote:
>> Try explaining to someone new that to use poll they'll have to do
>> their own memory management or artificially limit the number of
>> descriptors they'll use.
>
> My problem with poll() is this:

[...]

> I'd rather stick to select()

I believe I have edited you text suitably to make the actual content
visible. But this is a statement of nostalgic, emotional affection
and somewhat misplaced in a technical discussion.

From: Rainer Weikusat on
Rainer Weikusat <rweikusat(a)mssgmbh.com> writes:

> "Ersek, Laszlo" <lacos(a)caesar.elte.hu> writes:
>> On Tue, 4 May 2010, boltar2003(a)boltar.world wrote:
>>> Try explaining to someone new that to use poll they'll have to do
>>> their own memory management or artificially limit the number of
>>> descriptors they'll use.
>>
>> My problem with poll() is this:
>
> [...]
>
>> I'd rather stick to select()
>
> I believe I have edited you text suitably [...]

Accidental omission: The 'rather' was added by me.

From: Ersek, Laszlo on
On Tue, 4 May 2010, Rainer Weikusat wrote:

> "Ersek, Laszlo" <lacos(a)caesar.elte.hu> writes:

>> On Tue, 4 May 2010, boltar2003(a)boltar.world wrote:

>>> Try explaining to someone new that to use poll they'll have to do
>>> their own memory management or artificially limit the number of
>>> descriptors they'll use.
>>
>> My problem with poll() is this:
>
> [...]
>
>> I'd rather stick to select()
>
> I believe I have edited you text suitably to make the actual content
> visible.

Thank you for your effort.


> But this is a statement of nostalgic, emotional affection and somewhat
> misplaced in a technical discussion.

You could have tried to point out how to use the events / revents masks
for TCP, wrt. EOF, half-closed sockets, and urgent data. Or how to use
poll() correctly in combination with changing signal masks. Or what
POLLxxxx flags can be portably assumed equivalent in addition to what the
standard guarantees, and which POLLxxxx events can be ignored for TCP. If
these are obvious, you could have cited references.

BTW, your comment, attempting solely to deride me, is somewhat misplaced
in a technical discussion. You could have ignored my post, for example.

Have a nice day,
lacos
From: Rainer Weikusat on
"Ersek, Laszlo" <lacos(a)caesar.elte.hu> writes:
> On Tue, 4 May 2010, Rainer Weikusat wrote:
>
>> "Ersek, Laszlo" <lacos(a)caesar.elte.hu> writes:
>
>>> On Tue, 4 May 2010, boltar2003(a)boltar.world wrote:
>
>>>> Try explaining to someone new that to use poll they'll have to do
>>>> their own memory management or artificially limit the number of
>>>> descriptors they'll use.
>>>
>>> My problem with poll() is this:
>>
>> [...]
>>
>>> I'd rather stick to select()
>>
>> I believe I have edited you text suitably to make the actual content
>> visible. But this is a statement of nostalgic, emotional affection
>> and somewhat misplaced in a technical discussion.
>
> You could have tried to point out how to use the events / revents
> masks for TCP, wrt. EOF, half-closed sockets, and urgent data. Or how
> to use poll() correctly in combination with changing signal masks. Or
> what POLLxxxx flags can be portably assumed equivalent in addition to
> what the standard guarantees, and which POLLxxxx events can be ignored
> for TCP. If these are obvious, you could have cited references.


> BTW, your comment, attempting solely to deride me, is somewhat
> misplaced in a technical discussion.

I am honestly convinced that you (and the Boltar-person) are trying to
make up complicated problems based on hypothetical possibilities where
none actually exist in order to distract from the discussion of the
properties of different kinds of 'synchronous I/O multiplexing
interfaces' whose contents would otherwise not be completely favorable
regarding a solution both of you happen to prefer. And my 'comment'
attempts to communicate this in a succinct way instead falling into
one of your cleverly conceived 'bottomless side-issue trapdoors'.
From: Ersek, Laszlo on
On Tue, 4 May 2010, Rainer Weikusat wrote:

> I am honestly convinced that you (and the Boltar-person) are trying to
> make up complicated problems

I can't prove I do *not* want to do something.


> based on hypothetical possibilities where none actually exist in order
> to distract from the discussion

Why would I want to distract?


> of the properties of different kinds of 'synchronous I/O multiplexing
> interfaces' whose contents would otherwise not be completely favorable
> regarding a solution both of you happen to prefer.

Oh. So you think I'm trying to derail the discussion with personal
anecdotes because I'm supposedly afraid you might be able to prove poll()
superior, and then my life would have been in vain, writing all those
correct but suboptimally performing select() loops!

I see your point. It's wrong. I like to understand function specifications
in depth before calling said functions. No matter how superior you prove
poll() if I'm not sure I can call it correctly. When I say "I'd stick to
select() [unless you prove poll() un-messy -- please do!]", isn't that
invitation for you to prove poll() un-messy? I'm setting up my argument
for refutation by stating it as a fact. That should elicit more responses
from subscribers than asking questions. (Which I also did ask.) Instead of
disproving my post by describing poll()'s portably correct usage, you
accuse me of malice.


> And my 'comment' attempts to communicate this in a succinct way instead
> falling into one of your cleverly conceived 'bottomless side-issue
> trapdoors'.

I'm flattered.

(Though a bit dismayed seeing my evil plan laid bare.)

Cheers,
lacos