[Twisted-Python] reactor.callLater(0,) should not be an idiom[PATCH]

Bruce Mitchener bruce at cubik.org
Tue Oct 15 23:32:45 MDT 2002


Hello Moshe,

Moshe Zadka wrote:
> On Tue, 15 Oct 2002, "Sean Riley" <sean at twistedmatrix.com> wrote:
>>That doesn't answer my question. Shouldn't schedule() and callLater() appear
>>to the user to be called on the same module regardless of the underlying
>>implementation details?
> 
> That's not the correct question. The question is "is it possible?".
> The answer is that it is not possible without doing really obscure
> things. Thus, it shouldn't be done. If you'll recall, this was why
> .utils got there in the first place (namely, the process stuff).

Having an obscure and less usable API with some parts of the API on one 
object and other parts on some other object (or maybe even multiple 
other objects) isn't something that creates a nice usable and easy to 
learn/use framework.  Implementation details needn't (and shouldn't) 
bleed through into the public interfaces in the sort of manner that you 
are suggesting.  If they do, then perhaps something larger is wrong.

> Also, could you possibly observe common netiquette and stop quoting
> whole messages or putting your response before what you're responding
> to? Thank you.

You might try lightening your tone on the list.  You might be (are 
supposedly from what people have told me when I've asked) a decent 
person in real life, but you come across on this list as being overly 
confrontational, rude, and abrupt.

If 1.0 is happening soon, I'd hope that people asking (reasonable!) 
questions like this would be treated with some courtesy and respect and 
actual answers rather than the 'No. reactor is a "virtual" module.' sort 
of answer that provides no useful information.  However good Twisted may 
be, the attitudes that we show on the list when dealing with any user 
can be enough to convince people that Twisted is good or that it is bad 
and to be avoided.  I certainly hope that the latter doesn't (continue 
to) happen.

Regards,

  - Bruce





More information about the Twisted-Python mailing list