Skip to main content


Eclipse Community Forums

      Home
Home » Archived » Riena » Riena Manifest extensions
Riena Manifest extensions [message #4700] Fri, 13 June 2008 12:54 Go to next message
Eclipse UserFriend
Originally posted by: ekkehard.gentz-software.de

while analyzing the Riena Bundles I found in some MANIFEST.MF:

Riena-ForceStart: true

what does this mean ?

thx for info

ekke
Re: Riena Manifest extensions [message #4978 is a reply to message #4700] Tue, 17 June 2008 14:21 Go to previous messageGo to next message
Eclipse UserFriend
Riena-ForceStart is something proprietory that we invented and are not really happy about it. However consider this.
Your client has a dependency to riena.core and maybe riena.communication.core. Now it tries to start a remote service
proxy for the protocol proxy. You will probably know that you need riena.factory.hessian for this. However even if you
include riena.factory.hessian in the list of bundles for your started configuration, it will not get activated until a
class in there is accessed (thats what lazy loading does).

There is no way that we found so far to activate a bundle that your code has no direct or indirect dependency on. Code
that is in a bundle that creates an OSGi Service that you need indirect. Thats what Riena-ForceStart is for. It forces
all bundles that by its nature no one has a direct dependency on. riena.core does the check on Riena-ForceStart when
riena.core is activated.

Not sure if we will continue that way so dont rely on that we activate bundles with that header. :-)

christian

ekke schrieb:
> while analyzing the Riena Bundles I found in some MANIFEST.MF:
>
> Riena-ForceStart: true
>
> what does this mean ?
>
> thx for info
>
> ekke
Re: Riena Manifest extensions [message #7787 is a reply to message #4978] Tue, 19 August 2008 05:10 Go to previous messageGo to next message
Eclipse UserFriend
Autostarting bundles can also be achieved by setting
osgi.bundles property in your config.ini

This is to my knowledge the only "official" way to kickstart bundles and
bypassing the lazy loading scheme

Christian Campo kirjutas mulle midagi seesugust:
--
> Riena-ForceStart is something proprietory that we invented and are not
> really happy about it. However consider this. Your client has a
> dependency to riena.core and maybe riena.communication.core. Now it
> tries to start a remote service proxy for the protocol proxy. You will
> probably know that you need riena.factory.hessian for this. However even
> if you include riena.factory.hessian in the list of bundles for your
> started configuration, it will not get activated until a class in there
> is accessed (thats what lazy loading does).
>
> There is no way that we found so far to activate a bundle that your code
> has no direct or indirect dependency on. Code that is in a bundle that
> creates an OSGi Service that you need indirect. Thats what
> Riena-ForceStart is for. It forces all bundles that by its nature no one
> has a direct dependency on. riena.core does the check on
> Riena-ForceStart when riena.core is activated.
>
> Not sure if we will continue that way so dont rely on that we activate
> bundles with that header. :-)
>
> christian
>
> ekke schrieb:
>> while analyzing the Riena Bundles I found in some MANIFEST.MF:
>>
>> Riena-ForceStart: true
>>
>> what does this mean ?
>>
>> thx for info
>>
>> ekke

--
Roland Tepp
Re: Riena Manifest extensions [message #7808 is a reply to message #7787] Wed, 20 August 2008 07:56 Go to previous message
Eclipse UserFriend
Roland Tepp schrieb:
> Autostarting bundles can also be achieved by setting
> osgi.bundles property in your config.ini
>
> This is to my knowledge the only "official" way to kickstart bundles and
> bypassing the lazy loading scheme
>
> Christian Campo kirjutas mulle midagi seesugust:
right....using config.ini is a specified standard way.

However often you use a generated config.ini file, because its easier while you develop and add or remove bundles.
Config.ini are also under the control of the customer who sets up the finished software. So thats a little late if the
base framework that the customer's application relies on needs a certain bundle to be started. Then you want a
mechanisme to force the framework to work properly. Thats why we invented this property for the time being (until there
is a better approach)

christian
Re: Riena Manifest extensions [message #572124 is a reply to message #4700] Tue, 17 June 2008 14:21 Go to previous message
Eclipse UserFriend
Riena-ForceStart is something proprietory that we invented and are not really happy about it. However consider this.
Your client has a dependency to riena.core and maybe riena.communication.core. Now it tries to start a remote service
proxy for the protocol proxy. You will probably know that you need riena.factory.hessian for this. However even if you
include riena.factory.hessian in the list of bundles for your started configuration, it will not get activated until a
class in there is accessed (thats what lazy loading does).

There is no way that we found so far to activate a bundle that your code has no direct or indirect dependency on. Code
that is in a bundle that creates an OSGi Service that you need indirect. Thats what Riena-ForceStart is for. It forces
all bundles that by its nature no one has a direct dependency on. riena.core does the check on Riena-ForceStart when
riena.core is activated.

Not sure if we will continue that way so dont rely on that we activate bundles with that header. :-)

christian

ekke schrieb:
> while analyzing the Riena Bundles I found in some MANIFEST.MF:
>
> Riena-ForceStart: true
>
> what does this mean ?
>
> thx for info
>
> ekke
Re: Riena Manifest extensions [message #573331 is a reply to message #4978] Tue, 19 August 2008 05:10 Go to previous message
Eclipse UserFriend
Autostarting bundles can also be achieved by setting
osgi.bundles property in your config.ini

This is to my knowledge the only "official" way to kickstart bundles and
bypassing the lazy loading scheme

Christian Campo kirjutas mulle midagi seesugust:
--
> Riena-ForceStart is something proprietory that we invented and are not
> really happy about it. However consider this. Your client has a
> dependency to riena.core and maybe riena.communication.core. Now it
> tries to start a remote service proxy for the protocol proxy. You will
> probably know that you need riena.factory.hessian for this. However even
> if you include riena.factory.hessian in the list of bundles for your
> started configuration, it will not get activated until a class in there
> is accessed (thats what lazy loading does).
>
> There is no way that we found so far to activate a bundle that your code
> has no direct or indirect dependency on. Code that is in a bundle that
> creates an OSGi Service that you need indirect. Thats what
> Riena-ForceStart is for. It forces all bundles that by its nature no one
> has a direct dependency on. riena.core does the check on
> Riena-ForceStart when riena.core is activated.
>
> Not sure if we will continue that way so dont rely on that we activate
> bundles with that header. :-)
>
> christian
>
> ekke schrieb:
>> while analyzing the Riena Bundles I found in some MANIFEST.MF:
>>
>> Riena-ForceStart: true
>>
>> what does this mean ?
>>
>> thx for info
>>
>> ekke

--
Roland Tepp
Re: Riena Manifest extensions [message #573353 is a reply to message #7787] Wed, 20 August 2008 07:56 Go to previous message
Eclipse UserFriend
Roland Tepp schrieb:
> Autostarting bundles can also be achieved by setting
> osgi.bundles property in your config.ini
>
> This is to my knowledge the only "official" way to kickstart bundles and
> bypassing the lazy loading scheme
>
> Christian Campo kirjutas mulle midagi seesugust:
right....using config.ini is a specified standard way.

However often you use a generated config.ini file, because its easier while you develop and add or remove bundles.
Config.ini are also under the control of the customer who sets up the finished software. So thats a little late if the
base framework that the customer's application relies on needs a certain bundle to be started. Then you want a
mechanisme to force the framework to work properly. Thats why we invented this property for the time being (until there
is a better approach)

christian
Previous Topic:Milestone 1.0.0.M4 is coming end of august
Next Topic:Eclipse Magazin Artikel
Goto Forum:
  


Current Time: Fri Sep 19 07:25:55 EDT 2025

Powered by FUDForum. Page generated in 0.25234 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top

Sign up to our Newsletter

A fresh new issue delivered monthly