[Seaside-dev] [Dolphin] Use of Delay on CTPusher class>>startUp and pingProcess

[Julian Fitzell]

Hmm... but you understand why there *should* be a problem, right?

I mean I haven't tested this very thoroughly or even looked at the
code. I just got subclass should implement error when I updated Comet
before Pharo-Comet (which is the order we use everywhere else). I
don't know why it would have been working before and not now.

Is there something wrong with doing it when the server adaptor starts,
though? I think we should be avoiding class-side initialize methods
whenever possible anyway...

Julian

On Mon, Sep 21, 2009 at 8:45 AM, Lukas Renggli <renggli at gmail.com> wrote:
> I cannot really understand this, because nothing changed in the
> startup sequence. Also there are other Comet platform calls made
> before the one with the delay, so I don't see why it suddenly doesn't
> work anymore.
>
> Lukas
>
> 2009/9/21 Lukas Renggli <renggli at gmail.com>:
>> Ouch, I'll have a look into this.
>>
>> Thank you.
>>
>> Lukas
>>
>> 2009/9/21 Julian Fitzell <jfitzell at gmail.com>:
>>> That doesn't quite work. Seaside-Comet gets loaded before
>>> Seaside-Pharo-Comet and tries to use the platform method in an
>>> #initialize (just had this problem while loading). Can we have the
>>> server adaptor check if the process is running on startup?
>>>
>>> Julian
>>>
>>> On Sat, Sep 19, 2009 at 1:03 PM, Lukas Renggli <renggli at gmail.com> wrote:
>>>> Fixed. http://code.google.com/p/seaside/issues/detail?id=468
>>>>
>>>> 2009/9/19 Julian Fitzell <jfitzell at gmail.com>:
>>>>> On 9/18/09, Udo Schneider <Udo.Schneider at homeaddress.de> wrote:
>>>>>> Sebastian,
>>>>>>
>>>>>> > pingProcess
>>>>>> >        self allInstancesDo: [ :each | each ping ].
>>>>>> >        delay wait
>>>>>> >
>>>>>> > I think "delay wait" can be safely replaced by "Delay forSeconds: 15"
>>>>>> > (removing it from startUp).
>>>>>> >
>>>>>> >
>>>>>> You're right ... as long as you mean " (Delay forSeconds: 15) wait" :-)
>>>>>>
>>>>>> BTW: Could you replace "self allInstancesDo:" with "self allInstances do:"
>>>>>> ... at least in my image #allInstancesDo: is not defined.
>>>>>
>>>>> We need to be careful with doing anything with all instances actually
>>>>> because it's a terrible idea on GemStone. But I guess Comet depends on
>>>>> Comanche and GS is using Swazoo so it's probably ok in this case...
>>>>> still it can't be the most efficient implementation I would think.
>>>>>
>>>>> Julian
>>>>> _______________________________________________
>>>>> seaside-dev mailing list
>>>>> seaside-dev at lists.squeakfoundation.org
>>>>> http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Lukas Renggli
>>>> http://www.lukas-renggli.ch
>>>> _______________________________________________
>>>> seaside-dev mailing list
>>>> seaside-dev at lists.squeakfoundation.org
>>>> http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
>>>>
>>> _______________________________________________
>>> seaside-dev mailing list
>>> seaside-dev at lists.squeakfoundation.org
>>> http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
>>>
>>
>>
>>
>> --
>> Lukas Renggli
>> http://www.lukas-renggli.ch
>>
>
>
>
> --
> Lukas Renggli
> http://www.lukas-renggli.ch
> _______________________________________________
> seaside-dev mailing list
> seaside-dev at lists.squeakfoundation.org
> http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
>