You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@activemq.apache.org by Hiram Chirino <hi...@hiramchirino.com> on 2007/02/25 08:43:17 UTC
Do we really have to set a timestamp on the broker?
Rob, do you remember why added the following feature?
http://svn.apache.org/viewvc?view=rev&revision=376260
Setting the time on a message can break JMS spec compliance, plus it
forces us to have the overhead of getting the system time on every
message sent. I was hoping we could roll this one back.
--
Regards,
Hiram
Blog: http://hiramchirino.com
Re: Do we really have to set a timestamp on the broker?
Posted by Hiram Chirino <hi...@hiramchirino.com>.
Going to move this to a broker plugin for now then.
On 2/25/07, Rob Davies <ra...@gmail.com> wrote:
> exactly! but its sunday mornin :) - don't expect clarity of thought -
> yet!
>
> On 25 Feb 2007, at 08:04, Jason Dillon wrote:
>
> > Why not optional with a different property, so that the spec
> > behavior is always consistent, but users can still get the custom
> > data they need if the broker has some flag flipped on?
> >
> > --jason
> >
> >
> > On Feb 24, 2007, at 11:54 PM, Rob Davies wrote:
> >
> >> It was requested by a some users - they wanted to compare messages
> >> from different producers on different machines by time sent - and
> >> as the clients clocks aren't in sync, they really needed the
> >> timestamp set on the broker. I guess the best thing to do would to
> >> make it optional, off by default
> >>
> >> cheers,
> >>
> >> Rob
> >>
> >> On 25 Feb 2007, at 07:43, Hiram Chirino wrote:
> >>
> >>> Rob, do you remember why added the following feature?
> >>> http://svn.apache.org/viewvc?view=rev&revision=376260
> >>>
> >>> Setting the time on a message can break JMS spec compliance, plus it
> >>> forces us to have the overhead of getting the system time on every
> >>> message sent. I was hoping we could roll this one back.
> >>>
> >>> --
> >>> Regards,
> >>> Hiram
> >>>
> >>> Blog: http://hiramchirino.com
> >>
> >
>
>
--
Regards,
Hiram
Blog: http://hiramchirino.com
Re: Do we really have to set a timestamp on the broker?
Posted by Rob Davies <ra...@gmail.com>.
exactly! but its sunday mornin :) - don't expect clarity of thought -
yet!
On 25 Feb 2007, at 08:04, Jason Dillon wrote:
> Why not optional with a different property, so that the spec
> behavior is always consistent, but users can still get the custom
> data they need if the broker has some flag flipped on?
>
> --jason
>
>
> On Feb 24, 2007, at 11:54 PM, Rob Davies wrote:
>
>> It was requested by a some users - they wanted to compare messages
>> from different producers on different machines by time sent - and
>> as the clients clocks aren't in sync, they really needed the
>> timestamp set on the broker. I guess the best thing to do would to
>> make it optional, off by default
>>
>> cheers,
>>
>> Rob
>>
>> On 25 Feb 2007, at 07:43, Hiram Chirino wrote:
>>
>>> Rob, do you remember why added the following feature?
>>> http://svn.apache.org/viewvc?view=rev&revision=376260
>>>
>>> Setting the time on a message can break JMS spec compliance, plus it
>>> forces us to have the overhead of getting the system time on every
>>> message sent. I was hoping we could roll this one back.
>>>
>>> --
>>> Regards,
>>> Hiram
>>>
>>> Blog: http://hiramchirino.com
>>
>
Re: Do we really have to set a timestamp on the broker?
Posted by Jason Dillon <ja...@planet57.com>.
Is there a new snapshot out with this change?
--jason
On Feb 25, 2007, at 12:42 AM, Hiram Chirino wrote:
> Well,
>
> I've moved the existing broker time stamp logic to a
> TimeStampingBrokerPlugin. Anyone who depended on ActiveMQ 4.1.0's
> broker time stamping behavior will now have to enable the
> "timeStampingBrokerPlugin" in their broker configuration in subsequent
> releases.
>
> On 2/25/07, Jason Dillon <ja...@planet57.com> wrote:
>> Why not optional with a different property, so that the spec behavior
>> is always consistent, but users can still get the custom data they
>> need if the broker has some flag flipped on?
>
> This sounds like a cleaner approach, but since the behavior is
> different would need to be implemented a different plugin, or an
> option enabled on the TimeStampingBrokerPlugin.
>
>>
>> --jason
>>
>>
>> On Feb 24, 2007, at 11:54 PM, Rob Davies wrote:
>>
>> > It was requested by a some users - they wanted to compare messages
>> > from different producers on different machines by time sent - and
>> > as the clients clocks aren't in sync, they really needed the
>> > timestamp set on the broker. I guess the best thing to do would to
>> > make it optional, off by default
>> >
>> > cheers,
>> >
>> > Rob
>> >
>> > On 25 Feb 2007, at 07:43, Hiram Chirino wrote:
>> >
>> >> Rob, do you remember why added the following feature?
>> >> http://svn.apache.org/viewvc?view=rev&revision=376260
>> >>
>> >> Setting the time on a message can break JMS spec compliance,
>> plus it
>> >> forces us to have the overhead of getting the system time on every
>> >> message sent. I was hoping we could roll this one back.
>> >>
>> >> --
>> >> Regards,
>> >> Hiram
>> >>
>> >> Blog: http://hiramchirino.com
>> >
>>
>>
>
>
> --
> Regards,
> Hiram
>
> Blog: http://hiramchirino.com
Re: Do we really have to set a timestamp on the broker?
Posted by Hiram Chirino <hi...@hiramchirino.com>.
Well,
I've moved the existing broker time stamp logic to a
TimeStampingBrokerPlugin. Anyone who depended on ActiveMQ 4.1.0's
broker time stamping behavior will now have to enable the
"timeStampingBrokerPlugin" in their broker configuration in subsequent
releases.
On 2/25/07, Jason Dillon <ja...@planet57.com> wrote:
> Why not optional with a different property, so that the spec behavior
> is always consistent, but users can still get the custom data they
> need if the broker has some flag flipped on?
This sounds like a cleaner approach, but since the behavior is
different would need to be implemented a different plugin, or an
option enabled on the TimeStampingBrokerPlugin.
>
> --jason
>
>
> On Feb 24, 2007, at 11:54 PM, Rob Davies wrote:
>
> > It was requested by a some users - they wanted to compare messages
> > from different producers on different machines by time sent - and
> > as the clients clocks aren't in sync, they really needed the
> > timestamp set on the broker. I guess the best thing to do would to
> > make it optional, off by default
> >
> > cheers,
> >
> > Rob
> >
> > On 25 Feb 2007, at 07:43, Hiram Chirino wrote:
> >
> >> Rob, do you remember why added the following feature?
> >> http://svn.apache.org/viewvc?view=rev&revision=376260
> >>
> >> Setting the time on a message can break JMS spec compliance, plus it
> >> forces us to have the overhead of getting the system time on every
> >> message sent. I was hoping we could roll this one back.
> >>
> >> --
> >> Regards,
> >> Hiram
> >>
> >> Blog: http://hiramchirino.com
> >
>
>
--
Regards,
Hiram
Blog: http://hiramchirino.com
Re: Do we really have to set a timestamp on the broker?
Posted by Jason Dillon <ja...@planet57.com>.
Why not optional with a different property, so that the spec behavior
is always consistent, but users can still get the custom data they
need if the broker has some flag flipped on?
--jason
On Feb 24, 2007, at 11:54 PM, Rob Davies wrote:
> It was requested by a some users - they wanted to compare messages
> from different producers on different machines by time sent - and
> as the clients clocks aren't in sync, they really needed the
> timestamp set on the broker. I guess the best thing to do would to
> make it optional, off by default
>
> cheers,
>
> Rob
>
> On 25 Feb 2007, at 07:43, Hiram Chirino wrote:
>
>> Rob, do you remember why added the following feature?
>> http://svn.apache.org/viewvc?view=rev&revision=376260
>>
>> Setting the time on a message can break JMS spec compliance, plus it
>> forces us to have the overhead of getting the system time on every
>> message sent. I was hoping we could roll this one back.
>>
>> --
>> Regards,
>> Hiram
>>
>> Blog: http://hiramchirino.com
>
Re: Do we really have to set a timestamp on the broker?
Posted by Rob Davies <ra...@gmail.com>.
It was requested by a some users - they wanted to compare messages
from different producers on different machines by time sent - and as
the clients clocks aren't in sync, they really needed the timestamp
set on the broker. I guess the best thing to do would to make it
optional, off by default
cheers,
Rob
On 25 Feb 2007, at 07:43, Hiram Chirino wrote:
> Rob, do you remember why added the following feature?
> http://svn.apache.org/viewvc?view=rev&revision=376260
>
> Setting the time on a message can break JMS spec compliance, plus it
> forces us to have the overhead of getting the system time on every
> message sent. I was hoping we could roll this one back.
>
> --
> Regards,
> Hiram
>
> Blog: http://hiramchirino.com