RE: Tapestry Upgrade problem
We are going with the solution that I wrote about in my last post. After further investigation we believe that the PageSpecificationResolverImpl class is being initialized twice. We are not 100% certain of this, but unfortunately we don't have the time and resources to futher investigate this issue. We did discover that the synchronized call must be placed around the call to PageSpecificationResolver.resolve(...), trying to synchronize the method itself didn't work. Thanks for all your help Mike From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca] Sent: Mon 1/12/2009 9:53 AM To: Tapestry users Subject: RE: Tapestry Upgrade problem We believe that we may have found a solution to this problem. We have placed the call to PageSpecificationResolver.resolve(IRequestCycle cycle, String prefixedName) in PageSource.makeObject(Object key) in a synchronized block. With this in place we are unable to recreate the "no shell component found" problem we've been having. We intend to dig deeper and hopefully come up with a good explaination, (or theory at least) as to what caused the problem. Thanks Mike From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca] Sent: Fri 1/9/2009 11:36 AM To: Tapestry users Subject: RE: Tapestry Upgrade problem Done. If you need any more information please don't hesitate to ask. Thanks Mike From: andre...@gmail.com on behalf of Andreas Andreou Sent: Thu 1/8/2009 2:23 PM To: Tapestry users Subject: Re: Tapestry Upgrade problem Also, add an issue at https://issues.apache.org/jira/browse/TAPESTRY in order to track this and add more info, e.t.c. On Thu, Jan 8, 2009 at 9:03 PM, Andreas Andreou wrote: > I want to take a look, but got two questions... > - you said you tried several VMs, does that include 1.6 ones? > - Does https://issues.apache.org/jira/browse/TAPESTRY-2708 seem related? > > On Thu, Jan 8, 2009 at 8:18 PM, Mike Wasserman > wrote: >> We've tracked the problem further and it appears to be a thread safety >> problem in the PageLoader class. >> We have pop-ups in our app that are seperate Tapestry pages and we think >> that they may be clobbering each other when getting loaded. >> >> Is this possible? >> >> >> >> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca] >> Sent: Thu 1/8/2009 10:15 AM >> To: Tapestry users >> Subject: RE: Tapestry Upgrade problem >> >> >> >> It now makes sense to me why the _components map is empty when >> getComponent() is called. >> The map is never populated because no componentIds are found when >> PageLoder.construct() tries to get them off the IComponentSpecification, >> (PageLoader line: 358). >> The loop that in which the component creation happens is controlled by the >> number of componentIds found. >> >> Now to figure out why there are no ids. >> >> Mike >> >> >> >> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca] >> Sent: Wed 1/7/2009 3:32 PM >> To: Tapestry users >> Subject: RE: Tapestry Upgrade problem >> >> >> >> Actually I was wrong in what I wrote in my last email. >> addComponent() is NOT being called before getComponent() when the problem >> arises. >> >> Mike >> >> >> >> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca] >> Sent: Wed 1/7/2009 2:47 PM >> To: Tapestry users >> Subject: RE: Tapestry Upgrade problem >> >> >> >> Multiple pages are afflicted. It appears to happen primarily with the >> "shell" component. >> >> We put logging in the getComponent() and the addComponent() methods of the >> AbstractComponent class to output the Thread Id and Component Id since the >> adding of components and the getting of components seemed to be happening on >> the same thread. We got output that indicated that addComponent() was being >> called before getComponent(), but despite this we experienced the problem. >> When we stopped the app with a break point (we're using Eclipse) and >> examined the variables in the AbstractComponent class we found that the >> _components Map was null, and _active was false. >> >> Thanks >> >> Mike >> >> >> >> From: Howard Lewis Ship [mailto:hls...@gmail.com] >> Sent: Wed 1/7/2009 2:33 PM >> To: Tapestry users >> Subject: Re: Tapestry Upgrade problem >> >> >> >> The
RE: Tapestry Upgrade problem
We believe that we may have found a solution to this problem.
We have placed the call to PageSpecificationResolver.resolve(IRequestCycle
cycle, String prefixedName) in PageSource.makeObject(Object key) in a
synchronized block.
With this in place we are unable to recreate the "no shell component found"
problem we've been having.
We intend to dig deeper and hopefully come up with a good explaination, (or
theory at least) as to what caused the problem.
Thanks
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Fri 1/9/2009 11:36 AM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Done.
If you need any more information please don't hesitate to ask.
Thanks
Mike
From: andre...@gmail.com on behalf of Andreas Andreou
Sent: Thu 1/8/2009 2:23 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
Also, add an issue at https://issues.apache.org/jira/browse/TAPESTRY
in order to track this and add more info, e.t.c.
On Thu, Jan 8, 2009 at 9:03 PM, Andreas Andreou wrote:
> I want to take a look, but got two questions...
> - you said you tried several VMs, does that include 1.6 ones?
> - Does https://issues.apache.org/jira/browse/TAPESTRY-2708 seem related?
>
> On Thu, Jan 8, 2009 at 8:18 PM, Mike Wasserman
> wrote:
>> We've tracked the problem further and it appears to be a thread safety
>> problem in the PageLoader class.
>> We have pop-ups in our app that are seperate Tapestry pages and we think
>> that they may be clobbering each other when getting loaded.
>>
>> Is this possible?
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Thu 1/8/2009 10:15 AM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> It now makes sense to me why the _components map is empty when
>> getComponent() is called.
>> The map is never populated because no componentIds are found when
>> PageLoder.construct() tries to get them off the IComponentSpecification,
>> (PageLoader line: 358).
>> The loop that in which the component creation happens is controlled by the
>> number of componentIds found.
>>
>> Now to figure out why there are no ids.
>>
>> Mike
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Wed 1/7/2009 3:32 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> Actually I was wrong in what I wrote in my last email.
>> addComponent() is NOT being called before getComponent() when the problem
>> arises.
>>
>> Mike
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Wed 1/7/2009 2:47 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> Multiple pages are afflicted. It appears to happen primarily with the
>> "shell" component.
>>
>> We put logging in the getComponent() and the addComponent() methods of the
>> AbstractComponent class to output the Thread Id and Component Id since the
>> adding of components and the getting of components seemed to be happening on
>> the same thread. We got output that indicated that addComponent() was being
>> called before getComponent(), but despite this we experienced the problem.
>> When we stopped the app with a break point (we're using Eclipse) and
>> examined the variables in the AbstractComponent class we found that the
>> _components Map was null, and _active was false.
>>
>> Thanks
>>
>> Mike
>>
>>
>>
>> From: Howard Lewis Ship [mailto:hls...@gmail.com]
>> Sent: Wed 1/7/2009 2:33 PM
>> To: Tapestry users
>> Subject: Re: Tapestry Upgrade problem
>>
>>
>>
>> The _components field should be populated before your code gets invoked.
>>
>> Without looking at the code, it sounds like a Tapestry bug; but I
>> can't imagine how a bug like that could have slipped through for so
>> long that you're the only one who sees it.
>>
>> Does this happen to only a single page, or are multiple pages afflicted?
>>
>> On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
>> wrote:
>>> I added the following code to the AbstractComponent.getComponent(String id)
>>> method:
>>>
>>>while(_components == null) {
>>> try {
>>> System.out.pri
RE: Tapestry Upgrade problem
Done.
If you need any more information please don't hesitate to ask.
Thanks
Mike
From: andre...@gmail.com on behalf of Andreas Andreou
Sent: Thu 1/8/2009 2:23 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
Also, add an issue at https://issues.apache.org/jira/browse/TAPESTRY
in order to track this and add more info, e.t.c.
On Thu, Jan 8, 2009 at 9:03 PM, Andreas Andreou wrote:
> I want to take a look, but got two questions...
> - you said you tried several VMs, does that include 1.6 ones?
> - Does https://issues.apache.org/jira/browse/TAPESTRY-2708 seem related?
>
> On Thu, Jan 8, 2009 at 8:18 PM, Mike Wasserman
> wrote:
>> We've tracked the problem further and it appears to be a thread safety
>> problem in the PageLoader class.
>> We have pop-ups in our app that are seperate Tapestry pages and we think
>> that they may be clobbering each other when getting loaded.
>>
>> Is this possible?
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Thu 1/8/2009 10:15 AM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> It now makes sense to me why the _components map is empty when
>> getComponent() is called.
>> The map is never populated because no componentIds are found when
>> PageLoder.construct() tries to get them off the IComponentSpecification,
>> (PageLoader line: 358).
>> The loop that in which the component creation happens is controlled by the
>> number of componentIds found.
>>
>> Now to figure out why there are no ids.
>>
>> Mike
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Wed 1/7/2009 3:32 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> Actually I was wrong in what I wrote in my last email.
>> addComponent() is NOT being called before getComponent() when the problem
>> arises.
>>
>> Mike
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Wed 1/7/2009 2:47 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> Multiple pages are afflicted. It appears to happen primarily with the
>> "shell" component.
>>
>> We put logging in the getComponent() and the addComponent() methods of the
>> AbstractComponent class to output the Thread Id and Component Id since the
>> adding of components and the getting of components seemed to be happening on
>> the same thread. We got output that indicated that addComponent() was being
>> called before getComponent(), but despite this we experienced the problem.
>> When we stopped the app with a break point (we're using Eclipse) and
>> examined the variables in the AbstractComponent class we found that the
>> _components Map was null, and _active was false.
>>
>> Thanks
>>
>> Mike
>>
>>
>>
>> From: Howard Lewis Ship [mailto:hls...@gmail.com]
>> Sent: Wed 1/7/2009 2:33 PM
>> To: Tapestry users
>> Subject: Re: Tapestry Upgrade problem
>>
>>
>>
>> The _components field should be populated before your code gets invoked.
>>
>> Without looking at the code, it sounds like a Tapestry bug; but I
>> can't imagine how a bug like that could have slipped through for so
>> long that you're the only one who sees it.
>>
>> Does this happen to only a single page, or are multiple pages afflicted?
>>
>> On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
>> wrote:
>>> I added the following code to the AbstractComponent.getComponent(String id)
>>> method:
>>>
>>>while(_components == null) {
>>> try {
>>> System.out.println("SLEEPING");
>>>Thread.currentThread().sleep(500);
>>> } catch (InterruptedException e) {
>>>e.printStackTrace();
>>> }
>>>}
>>>
>>> I was hoping that by having the app sleep for a short period of time the
>>> _components HashMap would eventually be populated.
>>> It seems to work for a while, but eventually the app ends up in a state of
>>> permanently sleeping (this seems to happen randomly, sometimes after 2
>>> tests, sometimes after 20 tests, sometimes after 50 tests).
>>> So it appears that the _components HashMap is
RE: Tapestry Upgrade problem
We haven't tried 1.6 and unfortunately it's not really an option for us due to
restrictions placed on us by the client.
It doesn't seem related to https://issues.apache.org/jira/browse/TAPESTRY-2708
. We aren't getting a NullPointerException and although that problems seems to
be in a similar area of the code it isn't the same stack trace that we're
getting.
Mike
From: andre...@gmail.com on behalf of Andreas Andreou
Sent: Thu 1/8/2009 2:03 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
I want to take a look, but got two questions...
- you said you tried several VMs, does that include 1.6 ones?
- Does https://issues.apache.org/jira/browse/TAPESTRY-2708 seem related?
On Thu, Jan 8, 2009 at 8:18 PM, Mike Wasserman
wrote:
> We've tracked the problem further and it appears to be a thread safety
> problem in the PageLoader class.
> We have pop-ups in our app that are seperate Tapestry pages and we think that
> they may be clobbering each other when getting loaded.
>
> Is this possible?
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 1/8/2009 10:15 AM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> It now makes sense to me why the _components map is empty when getComponent()
> is called.
> The map is never populated because no componentIds are found when
> PageLoder.construct() tries to get them off the IComponentSpecification,
> (PageLoader line: 358).
> The loop that in which the component creation happens is controlled by the
> number of componentIds found.
>
> Now to figure out why there are no ids.
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Wed 1/7/2009 3:32 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Actually I was wrong in what I wrote in my last email.
> addComponent() is NOT being called before getComponent() when the problem
> arises.
>
> Mike
>
> ____________
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Wed 1/7/2009 2:47 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Multiple pages are afflicted. It appears to happen primarily with the "shell"
> component.
>
> We put logging in the getComponent() and the addComponent() methods of the
> AbstractComponent class to output the Thread Id and Component Id since the
> adding of components and the getting of components seemed to be happening on
> the same thread. We got output that indicated that addComponent() was being
> called before getComponent(), but despite this we experienced the problem.
> When we stopped the app with a break point (we're using Eclipse) and examined
> the variables in the AbstractComponent class we found that the _components
> Map was null, and _active was false.
>
> Thanks
>
> Mike
>
>
>
> From: Howard Lewis Ship [mailto:hls...@gmail.com]
> Sent: Wed 1/7/2009 2:33 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> The _components field should be populated before your code gets invoked.
>
> Without looking at the code, it sounds like a Tapestry bug; but I
> can't imagine how a bug like that could have slipped through for so
> long that you're the only one who sees it.
>
> Does this happen to only a single page, or are multiple pages afflicted?
>
> On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
> wrote:
>> I added the following code to the AbstractComponent.getComponent(String id)
>> method:
>>
>>while(_components == null) {
>> try {
>> System.out.println("SLEEPING");
>>Thread.currentThread().sleep(500);
>> } catch (InterruptedException e) {
>>e.printStackTrace();
>> }
>>}
>>
>> I was hoping that by having the app sleep for a short period of time the
>> _components HashMap would eventually be populated.
>> It seems to work for a while, but eventually the app ends up in a state of
>> permanently sleeping (this seems to happen randomly, sometimes after 2
>> tests, sometimes after 20 tests, sometimes after 50 tests).
>> So it appears that the _components HashMap is never populated under some
>> circumstances.
>>
>> Anyone have any clue why this is happening, or how it is possible for this
>> to happen in Tapestry?
>>
>> Thanks
>>
>> Mike
>>
>>
>>
>>
>>
Re: Tapestry Upgrade problem
Sounds like you may be on to something!
On Thu, Jan 8, 2009 at 10:18 AM, Mike Wasserman
wrote:
> We've tracked the problem further and it appears to be a thread safety
> problem in the PageLoader class.
> We have pop-ups in our app that are seperate Tapestry pages and we think that
> they may be clobbering each other when getting loaded.
>
> Is this possible?
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 1/8/2009 10:15 AM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> It now makes sense to me why the _components map is empty when getComponent()
> is called.
> The map is never populated because no componentIds are found when
> PageLoder.construct() tries to get them off the IComponentSpecification,
> (PageLoader line: 358).
> The loop that in which the component creation happens is controlled by the
> number of componentIds found.
>
> Now to figure out why there are no ids.
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Wed 1/7/2009 3:32 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Actually I was wrong in what I wrote in my last email.
> addComponent() is NOT being called before getComponent() when the problem
> arises.
>
> Mike
>
> ________
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Wed 1/7/2009 2:47 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Multiple pages are afflicted. It appears to happen primarily with the "shell"
> component.
>
> We put logging in the getComponent() and the addComponent() methods of the
> AbstractComponent class to output the Thread Id and Component Id since the
> adding of components and the getting of components seemed to be happening on
> the same thread. We got output that indicated that addComponent() was being
> called before getComponent(), but despite this we experienced the problem.
> When we stopped the app with a break point (we're using Eclipse) and examined
> the variables in the AbstractComponent class we found that the _components
> Map was null, and _active was false.
>
> Thanks
>
> Mike
>
>
>
> From: Howard Lewis Ship [mailto:hls...@gmail.com]
> Sent: Wed 1/7/2009 2:33 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> The _components field should be populated before your code gets invoked.
>
> Without looking at the code, it sounds like a Tapestry bug; but I
> can't imagine how a bug like that could have slipped through for so
> long that you're the only one who sees it.
>
> Does this happen to only a single page, or are multiple pages afflicted?
>
> On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
> wrote:
>> I added the following code to the AbstractComponent.getComponent(String id)
>> method:
>>
>>while(_components == null) {
>> try {
>> System.out.println("SLEEPING");
>>Thread.currentThread().sleep(500);
>> } catch (InterruptedException e) {
>>e.printStackTrace();
>> }
>>}
>>
>> I was hoping that by having the app sleep for a short period of time the
>> _components HashMap would eventually be populated.
>> It seems to work for a while, but eventually the app ends up in a state of
>> permanently sleeping (this seems to happen randomly, sometimes after 2
>> tests, sometimes after 20 tests, sometimes after 50 tests).
>> So it appears that the _components HashMap is never populated under some
>> circumstances.
>>
>> Anyone have any clue why this is happening, or how it is possible for this
>> to happen in Tapestry?
>>
>> Thanks
>>
>> Mike
>>
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Tue 1/6/2009 2:58 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> I've tried a variety of JDKs and the problem happens with all of them.
>>
>> I've tracked it through the code and what it seems to come down to is that
>> sometimes the _components Map in the AbstractComponent class is populated
>> and other times it's null.
>> My guess at this point is that it's some sort of threading problem within
>> Tapestry.
>>
>> Anyone have any ideas on how to fix this problem?
>>
>> Thanks
>>
>>
Re: Tapestry Upgrade problem
Also, add an issue at https://issues.apache.org/jira/browse/TAPESTRY
in order to track this and add more info, e.t.c.
On Thu, Jan 8, 2009 at 9:03 PM, Andreas Andreou wrote:
> I want to take a look, but got two questions...
> - you said you tried several VMs, does that include 1.6 ones?
> - Does https://issues.apache.org/jira/browse/TAPESTRY-2708 seem related?
>
> On Thu, Jan 8, 2009 at 8:18 PM, Mike Wasserman
> wrote:
>> We've tracked the problem further and it appears to be a thread safety
>> problem in the PageLoader class.
>> We have pop-ups in our app that are seperate Tapestry pages and we think
>> that they may be clobbering each other when getting loaded.
>>
>> Is this possible?
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Thu 1/8/2009 10:15 AM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> It now makes sense to me why the _components map is empty when
>> getComponent() is called.
>> The map is never populated because no componentIds are found when
>> PageLoder.construct() tries to get them off the IComponentSpecification,
>> (PageLoader line: 358).
>> The loop that in which the component creation happens is controlled by the
>> number of componentIds found.
>>
>> Now to figure out why there are no ids.
>>
>> Mike
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Wed 1/7/2009 3:32 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> Actually I was wrong in what I wrote in my last email.
>> addComponent() is NOT being called before getComponent() when the problem
>> arises.
>>
>> Mike
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Wed 1/7/2009 2:47 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> Multiple pages are afflicted. It appears to happen primarily with the
>> "shell" component.
>>
>> We put logging in the getComponent() and the addComponent() methods of the
>> AbstractComponent class to output the Thread Id and Component Id since the
>> adding of components and the getting of components seemed to be happening on
>> the same thread. We got output that indicated that addComponent() was being
>> called before getComponent(), but despite this we experienced the problem.
>> When we stopped the app with a break point (we're using Eclipse) and
>> examined the variables in the AbstractComponent class we found that the
>> _components Map was null, and _active was false.
>>
>> Thanks
>>
>> Mike
>>
>>
>>
>> From: Howard Lewis Ship [mailto:hls...@gmail.com]
>> Sent: Wed 1/7/2009 2:33 PM
>> To: Tapestry users
>> Subject: Re: Tapestry Upgrade problem
>>
>>
>>
>> The _components field should be populated before your code gets invoked.
>>
>> Without looking at the code, it sounds like a Tapestry bug; but I
>> can't imagine how a bug like that could have slipped through for so
>> long that you're the only one who sees it.
>>
>> Does this happen to only a single page, or are multiple pages afflicted?
>>
>> On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
>> wrote:
>>> I added the following code to the AbstractComponent.getComponent(String id)
>>> method:
>>>
>>>while(_components == null) {
>>> try {
>>> System.out.println("SLEEPING");
>>>Thread.currentThread().sleep(500);
>>> } catch (InterruptedException e) {
>>>e.printStackTrace();
>>> }
>>>}
>>>
>>> I was hoping that by having the app sleep for a short period of time the
>>> _components HashMap would eventually be populated.
>>> It seems to work for a while, but eventually the app ends up in a state of
>>> permanently sleeping (this seems to happen randomly, sometimes after 2
>>> tests, sometimes after 20 tests, sometimes after 50 tests).
>>> So it appears that the _components HashMap is never populated under some
>>> circumstances.
>>>
>>> Anyone have any clue why this is happening, or how it is possible for this
>>> to happen in Tapestry?
>>>
>>> Thanks
>>>
>>> Mike
>
Re: Tapestry Upgrade problem
I want to take a look, but got two questions...
- you said you tried several VMs, does that include 1.6 ones?
- Does https://issues.apache.org/jira/browse/TAPESTRY-2708 seem related?
On Thu, Jan 8, 2009 at 8:18 PM, Mike Wasserman
wrote:
> We've tracked the problem further and it appears to be a thread safety
> problem in the PageLoader class.
> We have pop-ups in our app that are seperate Tapestry pages and we think that
> they may be clobbering each other when getting loaded.
>
> Is this possible?
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 1/8/2009 10:15 AM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> It now makes sense to me why the _components map is empty when getComponent()
> is called.
> The map is never populated because no componentIds are found when
> PageLoder.construct() tries to get them off the IComponentSpecification,
> (PageLoader line: 358).
> The loop that in which the component creation happens is controlled by the
> number of componentIds found.
>
> Now to figure out why there are no ids.
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Wed 1/7/2009 3:32 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Actually I was wrong in what I wrote in my last email.
> addComponent() is NOT being called before getComponent() when the problem
> arises.
>
> Mike
>
> ________
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Wed 1/7/2009 2:47 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Multiple pages are afflicted. It appears to happen primarily with the "shell"
> component.
>
> We put logging in the getComponent() and the addComponent() methods of the
> AbstractComponent class to output the Thread Id and Component Id since the
> adding of components and the getting of components seemed to be happening on
> the same thread. We got output that indicated that addComponent() was being
> called before getComponent(), but despite this we experienced the problem.
> When we stopped the app with a break point (we're using Eclipse) and examined
> the variables in the AbstractComponent class we found that the _components
> Map was null, and _active was false.
>
> Thanks
>
> Mike
>
>
>
> From: Howard Lewis Ship [mailto:hls...@gmail.com]
> Sent: Wed 1/7/2009 2:33 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> The _components field should be populated before your code gets invoked.
>
> Without looking at the code, it sounds like a Tapestry bug; but I
> can't imagine how a bug like that could have slipped through for so
> long that you're the only one who sees it.
>
> Does this happen to only a single page, or are multiple pages afflicted?
>
> On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
> wrote:
>> I added the following code to the AbstractComponent.getComponent(String id)
>> method:
>>
>>while(_components == null) {
>> try {
>> System.out.println("SLEEPING");
>>Thread.currentThread().sleep(500);
>> } catch (InterruptedException e) {
>>e.printStackTrace();
>> }
>>}
>>
>> I was hoping that by having the app sleep for a short period of time the
>> _components HashMap would eventually be populated.
>> It seems to work for a while, but eventually the app ends up in a state of
>> permanently sleeping (this seems to happen randomly, sometimes after 2
>> tests, sometimes after 20 tests, sometimes after 50 tests).
>> So it appears that the _components HashMap is never populated under some
>> circumstances.
>>
>> Anyone have any clue why this is happening, or how it is possible for this
>> to happen in Tapestry?
>>
>> Thanks
>>
>> Mike
>>
>>
>>
>>
>> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
>> Sent: Tue 1/6/2009 2:58 PM
>> To: Tapestry users
>> Subject: RE: Tapestry Upgrade problem
>>
>>
>>
>> I've tried a variety of JDKs and the problem happens with all of them.
>>
>> I've tracked it through the code and what it seems to come down to is that
>> sometimes the _components Map in the AbstractComponent class is populated
>> and other times it's null.
>> My guess at this point is that it's some sort of threading problem
RE: Tapestry Upgrade problem
We've tracked the problem further and it appears to be a thread safety problem
in the PageLoader class.
We have pop-ups in our app that are seperate Tapestry pages and we think that
they may be clobbering each other when getting loaded.
Is this possible?
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Thu 1/8/2009 10:15 AM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
It now makes sense to me why the _components map is empty when getComponent()
is called.
The map is never populated because no componentIds are found when
PageLoder.construct() tries to get them off the IComponentSpecification,
(PageLoader line: 358).
The loop that in which the component creation happens is controlled by the
number of componentIds found.
Now to figure out why there are no ids.
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Wed 1/7/2009 3:32 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Actually I was wrong in what I wrote in my last email.
addComponent() is NOT being called before getComponent() when the problem
arises.
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Wed 1/7/2009 2:47 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Multiple pages are afflicted. It appears to happen primarily with the "shell"
component.
We put logging in the getComponent() and the addComponent() methods of the
AbstractComponent class to output the Thread Id and Component Id since the
adding of components and the getting of components seemed to be happening on
the same thread. We got output that indicated that addComponent() was being
called before getComponent(), but despite this we experienced the problem.
When we stopped the app with a break point (we're using Eclipse) and examined
the variables in the AbstractComponent class we found that the _components Map
was null, and _active was false.
Thanks
Mike
From: Howard Lewis Ship [mailto:hls...@gmail.com]
Sent: Wed 1/7/2009 2:33 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
The _components field should be populated before your code gets invoked.
Without looking at the code, it sounds like a Tapestry bug; but I
can't imagine how a bug like that could have slipped through for so
long that you're the only one who sees it.
Does this happen to only a single page, or are multiple pages afflicted?
On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
wrote:
> I added the following code to the AbstractComponent.getComponent(String id)
> method:
>
>while(_components == null) {
> try {
> System.out.println("SLEEPING");
>Thread.currentThread().sleep(500);
> } catch (InterruptedException e) {
>e.printStackTrace();
> }
>}
>
> I was hoping that by having the app sleep for a short period of time the
> _components HashMap would eventually be populated.
> It seems to work for a while, but eventually the app ends up in a state of
> permanently sleeping (this seems to happen randomly, sometimes after 2 tests,
> sometimes after 20 tests, sometimes after 50 tests).
> So it appears that the _components HashMap is never populated under some
> circumstances.
>
> Anyone have any clue why this is happening, or how it is possible for this to
> happen in Tapestry?
>
> Thanks
>
> Mike
>
>
> ____
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Tue 1/6/2009 2:58 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> I've tried a variety of JDKs and the problem happens with all of them.
>
> I've tracked it through the code and what it seems to come down to is that
> sometimes the _components Map in the AbstractComponent class is populated and
> other times it's null.
> My guess at this point is that it's some sort of threading problem within
> Tapestry.
>
> Anyone have any ideas on how to fix this problem?
>
> Thanks
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 12/25/2008 1:07 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Thanks for your response.
> No, we didn't have this problem with 4.1.2
>
> I am off on holidays for a little while, but I will try your suggestion as
> soon as I get back.
>
> Thanks again
>
>
>
> From: andre...@gmail.com on behalf of Andreas Andreou
> Sent: Wed 12/24/2008 3:11 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> So, you weren't see
RE: Tapestry Upgrade problem
It now makes sense to me why the _components map is empty when getComponent()
is called.
The map is never populated because no componentIds are found when
PageLoder.construct() tries to get them off the IComponentSpecification,
(PageLoader line: 358).
The loop that in which the component creation happens is controlled by the
number of componentIds found.
Now to figure out why there are no ids.
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Wed 1/7/2009 3:32 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Actually I was wrong in what I wrote in my last email.
addComponent() is NOT being called before getComponent() when the problem
arises.
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Wed 1/7/2009 2:47 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Multiple pages are afflicted. It appears to happen primarily with the "shell"
component.
We put logging in the getComponent() and the addComponent() methods of the
AbstractComponent class to output the Thread Id and Component Id since the
adding of components and the getting of components seemed to be happening on
the same thread. We got output that indicated that addComponent() was being
called before getComponent(), but despite this we experienced the problem.
When we stopped the app with a break point (we're using Eclipse) and examined
the variables in the AbstractComponent class we found that the _components Map
was null, and _active was false.
Thanks
Mike
From: Howard Lewis Ship [mailto:hls...@gmail.com]
Sent: Wed 1/7/2009 2:33 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
The _components field should be populated before your code gets invoked.
Without looking at the code, it sounds like a Tapestry bug; but I
can't imagine how a bug like that could have slipped through for so
long that you're the only one who sees it.
Does this happen to only a single page, or are multiple pages afflicted?
On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
wrote:
> I added the following code to the AbstractComponent.getComponent(String id)
> method:
>
>while(_components == null) {
> try {
> System.out.println("SLEEPING");
>Thread.currentThread().sleep(500);
> } catch (InterruptedException e) {
>e.printStackTrace();
> }
>}
>
> I was hoping that by having the app sleep for a short period of time the
> _components HashMap would eventually be populated.
> It seems to work for a while, but eventually the app ends up in a state of
> permanently sleeping (this seems to happen randomly, sometimes after 2 tests,
> sometimes after 20 tests, sometimes after 50 tests).
> So it appears that the _components HashMap is never populated under some
> circumstances.
>
> Anyone have any clue why this is happening, or how it is possible for this to
> happen in Tapestry?
>
> Thanks
>
> Mike
>
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Tue 1/6/2009 2:58 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> I've tried a variety of JDKs and the problem happens with all of them.
>
> I've tracked it through the code and what it seems to come down to is that
> sometimes the _components Map in the AbstractComponent class is populated and
> other times it's null.
> My guess at this point is that it's some sort of threading problem within
> Tapestry.
>
> Anyone have any ideas on how to fix this problem?
>
> Thanks
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 12/25/2008 1:07 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Thanks for your response.
> No, we didn't have this problem with 4.1.2
>
> I am off on holidays for a little while, but I will try your suggestion as
> soon as I get back.
>
> Thanks again
>
>
>
> From: andre...@gmail.com on behalf of Andreas Andreou
> Sent: Wed 12/24/2008 3:11 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> So, you weren't seeing this problem with 4.1.2 ?
>
> I'd be interesting if you could try with a 1.6 VM, but apart from that
> and since you're getting this consistently, i'm wondering if you can try this
> (which is based on some old reports I remember reading here):
> - Start your app and before running the tests, try to visit all the affected
> pages by hand (making sure they load fine) and only then fire the tests.
> D
RE: Tapestry Upgrade problem
Actually I was wrong in what I wrote in my last email.
addComponent() is NOT being called before getComponent() when the problem
arises.
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Wed 1/7/2009 2:47 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Multiple pages are afflicted. It appears to happen primarily with the "shell"
component.
We put logging in the getComponent() and the addComponent() methods of the
AbstractComponent class to output the Thread Id and Component Id since the
adding of components and the getting of components seemed to be happening on
the same thread. We got output that indicated that addComponent() was being
called before getComponent(), but despite this we experienced the problem.
When we stopped the app with a break point (we're using Eclipse) and examined
the variables in the AbstractComponent class we found that the _components Map
was null, and _active was false.
Thanks
Mike
From: Howard Lewis Ship [mailto:hls...@gmail.com]
Sent: Wed 1/7/2009 2:33 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
The _components field should be populated before your code gets invoked.
Without looking at the code, it sounds like a Tapestry bug; but I
can't imagine how a bug like that could have slipped through for so
long that you're the only one who sees it.
Does this happen to only a single page, or are multiple pages afflicted?
On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
wrote:
> I added the following code to the AbstractComponent.getComponent(String id)
> method:
>
>while(_components == null) {
> try {
> System.out.println("SLEEPING");
>Thread.currentThread().sleep(500);
> } catch (InterruptedException e) {
>e.printStackTrace();
> }
>}
>
> I was hoping that by having the app sleep for a short period of time the
> _components HashMap would eventually be populated.
> It seems to work for a while, but eventually the app ends up in a state of
> permanently sleeping (this seems to happen randomly, sometimes after 2 tests,
> sometimes after 20 tests, sometimes after 50 tests).
> So it appears that the _components HashMap is never populated under some
> circumstances.
>
> Anyone have any clue why this is happening, or how it is possible for this to
> happen in Tapestry?
>
> Thanks
>
> Mike
>
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Tue 1/6/2009 2:58 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> I've tried a variety of JDKs and the problem happens with all of them.
>
> I've tracked it through the code and what it seems to come down to is that
> sometimes the _components Map in the AbstractComponent class is populated and
> other times it's null.
> My guess at this point is that it's some sort of threading problem within
> Tapestry.
>
> Anyone have any ideas on how to fix this problem?
>
> Thanks
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 12/25/2008 1:07 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Thanks for your response.
> No, we didn't have this problem with 4.1.2
>
> I am off on holidays for a little while, but I will try your suggestion as
> soon as I get back.
>
> Thanks again
>
>
>
> From: andre...@gmail.com on behalf of Andreas Andreou
> Sent: Wed 12/24/2008 3:11 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> So, you weren't seeing this problem with 4.1.2 ?
>
> I'd be interesting if you could try with a 1.6 VM, but apart from that
> and since you're getting this consistently, i'm wondering if you can try this
> (which is based on some old reports I remember reading here):
> - Start your app and before running the tests, try to visit all the affected
> pages by hand (making sure they load fine) and only then fire the tests.
> Do you see problems after that?
>
> On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman
> wrote:
>> I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the
>> same problem with both.
>> The Selenium tests are hitting the app pretty hard and I'm wondering if
>> it's causing a threading problem within Tapestry.
>> Isn't the component creation multi-threaded?
>>
>>
>> -Original Message-
>> From: Howard Lewis Ship [mailto:hls...@gmail.com]
>> Sent: Tuesday, December 23, 2008 2:00 PM
RE: Tapestry Upgrade problem
Multiple pages are afflicted. It appears to happen primarily with the "shell"
component.
We put logging in the getComponent() and the addComponent() methods of the
AbstractComponent class to output the Thread Id and Component Id since the
adding of components and the getting of components seemed to be happening on
the same thread. We got output that indicated that addComponent() was being
called before getComponent(), but despite this we experienced the problem.
When we stopped the app with a break point (we're using Eclipse) and examined
the variables in the AbstractComponent class we found that the _components Map
was null, and _active was false.
Thanks
Mike
From: Howard Lewis Ship [mailto:hls...@gmail.com]
Sent: Wed 1/7/2009 2:33 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
The _components field should be populated before your code gets invoked.
Without looking at the code, it sounds like a Tapestry bug; but I
can't imagine how a bug like that could have slipped through for so
long that you're the only one who sees it.
Does this happen to only a single page, or are multiple pages afflicted?
On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
wrote:
> I added the following code to the AbstractComponent.getComponent(String id)
> method:
>
>while(_components == null) {
> try {
> System.out.println("SLEEPING");
>Thread.currentThread().sleep(500);
> } catch (InterruptedException e) {
>e.printStackTrace();
> }
>}
>
> I was hoping that by having the app sleep for a short period of time the
> _components HashMap would eventually be populated.
> It seems to work for a while, but eventually the app ends up in a state of
> permanently sleeping (this seems to happen randomly, sometimes after 2 tests,
> sometimes after 20 tests, sometimes after 50 tests).
> So it appears that the _components HashMap is never populated under some
> circumstances.
>
> Anyone have any clue why this is happening, or how it is possible for this to
> happen in Tapestry?
>
> Thanks
>
> Mike
>
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Tue 1/6/2009 2:58 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> I've tried a variety of JDKs and the problem happens with all of them.
>
> I've tracked it through the code and what it seems to come down to is that
> sometimes the _components Map in the AbstractComponent class is populated and
> other times it's null.
> My guess at this point is that it's some sort of threading problem within
> Tapestry.
>
> Anyone have any ideas on how to fix this problem?
>
> Thanks
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 12/25/2008 1:07 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Thanks for your response.
> No, we didn't have this problem with 4.1.2
>
> I am off on holidays for a little while, but I will try your suggestion as
> soon as I get back.
>
> Thanks again
>
>
>
> From: andre...@gmail.com on behalf of Andreas Andreou
> Sent: Wed 12/24/2008 3:11 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> So, you weren't seeing this problem with 4.1.2 ?
>
> I'd be interesting if you could try with a 1.6 VM, but apart from that
> and since you're getting this consistently, i'm wondering if you can try this
> (which is based on some old reports I remember reading here):
> - Start your app and before running the tests, try to visit all the affected
> pages by hand (making sure they load fine) and only then fire the tests.
> Do you see problems after that?
>
> On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman
> wrote:
>> I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the
>> same problem with both.
>> The Selenium tests are hitting the app pretty hard and I'm wondering if
>> it's causing a threading problem within Tapestry.
>> Isn't the component creation multi-threaded?
>>
>>
>> -Original Message-
>> From: Howard Lewis Ship [mailto:hls...@gmail.com]
>> Sent: Tuesday, December 23, 2008 2:00 PM
>> To: Tapestry users
>> Subject: Re: Tapestry Upgrade problem
>>
>> What JDK are you running on?
>>
>> On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman
>> wrote:
>>> I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6.
>>> When I run my integration tests (using
Re: Tapestry Upgrade problem
The _components field should be populated before your code gets invoked.
Without looking at the code, it sounds like a Tapestry bug; but I
can't imagine how a bug like that could have slipped through for so
long that you're the only one who sees it.
Does this happen to only a single page, or are multiple pages afflicted?
On Wed, Jan 7, 2009 at 10:40 AM, Mike Wasserman
wrote:
> I added the following code to the AbstractComponent.getComponent(String id)
> method:
>
>while(_components == null) {
> try {
> System.out.println("SLEEPING");
>Thread.currentThread().sleep(500);
> } catch (InterruptedException e) {
>e.printStackTrace();
> }
>}
>
> I was hoping that by having the app sleep for a short period of time the
> _components HashMap would eventually be populated.
> It seems to work for a while, but eventually the app ends up in a state of
> permanently sleeping (this seems to happen randomly, sometimes after 2 tests,
> sometimes after 20 tests, sometimes after 50 tests).
> So it appears that the _components HashMap is never populated under some
> circumstances.
>
> Anyone have any clue why this is happening, or how it is possible for this to
> happen in Tapestry?
>
> Thanks
>
> Mike
>
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Tue 1/6/2009 2:58 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> I've tried a variety of JDKs and the problem happens with all of them.
>
> I've tracked it through the code and what it seems to come down to is that
> sometimes the _components Map in the AbstractComponent class is populated and
> other times it's null.
> My guess at this point is that it's some sort of threading problem within
> Tapestry.
>
> Anyone have any ideas on how to fix this problem?
>
> Thanks
>
> Mike
>
>
>
> From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
> Sent: Thu 12/25/2008 1:07 PM
> To: Tapestry users
> Subject: RE: Tapestry Upgrade problem
>
>
>
> Thanks for your response.
> No, we didn't have this problem with 4.1.2
>
> I am off on holidays for a little while, but I will try your suggestion as
> soon as I get back.
>
> Thanks again
>
>
>
> From: andre...@gmail.com on behalf of Andreas Andreou
> Sent: Wed 12/24/2008 3:11 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
>
>
> So, you weren't seeing this problem with 4.1.2 ?
>
> I'd be interesting if you could try with a 1.6 VM, but apart from that
> and since you're getting this consistently, i'm wondering if you can try this
> (which is based on some old reports I remember reading here):
> - Start your app and before running the tests, try to visit all the affected
> pages by hand (making sure they load fine) and only then fire the tests.
> Do you see problems after that?
>
> On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman
> wrote:
>> I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the
>> same problem with both.
>> The Selenium tests are hitting the app pretty hard and I'm wondering if
>> it's causing a threading problem within Tapestry.
>> Isn't the component creation multi-threaded?
>>
>>
>> -Original Message-
>> From: Howard Lewis Ship [mailto:hls...@gmail.com]
>> Sent: Tuesday, December 23, 2008 2:00 PM
>> To: Tapestry users
>> Subject: Re: Tapestry Upgrade problem
>>
>> What JDK are you running on?
>>
>> On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman
>> wrote:
>>> I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6.
>>> When I run my integration tests (using Selenium) I am get a error
>> randomly.
>>> I receive a org.apache.hivemind.ApplicationRuntimeException that
>> indicates that given component (page) does not contain a component
>> shell.
>>> All of the affected pages have which leads me to
>> believe that I have specified a shell component.
>>> I have tracked the problem into the PageLoader class, but since the
>> problem is random it is very difficult to debug.
>>> Is it possible that this is a threading problem?
>>>
>>> Can anyone help?
>>>
>>> Mike
>>>
>>
>>
>>
>> --
>> Howard M. Lewis Ship
>>
>> Creator Apache Tapestry and Apache HiveMind
>>
>> -
>> To unsubscribe, e-ma
RE: Tapestry Upgrade problem
I added the following code to the AbstractComponent.getComponent(String id)
method:
while(_components == null) {
try {
System.out.println("SLEEPING");
Thread.currentThread().sleep(500);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
I was hoping that by having the app sleep for a short period of time the
_components HashMap would eventually be populated.
It seems to work for a while, but eventually the app ends up in a state of
permanently sleeping (this seems to happen randomly, sometimes after 2 tests,
sometimes after 20 tests, sometimes after 50 tests).
So it appears that the _components HashMap is never populated under some
circumstances.
Anyone have any clue why this is happening, or how it is possible for this to
happen in Tapestry?
Thanks
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Tue 1/6/2009 2:58 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
I've tried a variety of JDKs and the problem happens with all of them.
I've tracked it through the code and what it seems to come down to is that
sometimes the _components Map in the AbstractComponent class is populated and
other times it's null.
My guess at this point is that it's some sort of threading problem within
Tapestry.
Anyone have any ideas on how to fix this problem?
Thanks
Mike
From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca]
Sent: Thu 12/25/2008 1:07 PM
To: Tapestry users
Subject: RE: Tapestry Upgrade problem
Thanks for your response.
No, we didn't have this problem with 4.1.2
I am off on holidays for a little while, but I will try your suggestion as soon
as I get back.
Thanks again
From: andre...@gmail.com on behalf of Andreas Andreou
Sent: Wed 12/24/2008 3:11 PM
To: Tapestry users
Subject: Re: Tapestry Upgrade problem
So, you weren't seeing this problem with 4.1.2 ?
I'd be interesting if you could try with a 1.6 VM, but apart from that
and since you're getting this consistently, i'm wondering if you can try this
(which is based on some old reports I remember reading here):
- Start your app and before running the tests, try to visit all the affected
pages by hand (making sure they load fine) and only then fire the tests.
Do you see problems after that?
On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman
wrote:
> I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the
> same problem with both.
> The Selenium tests are hitting the app pretty hard and I'm wondering if
> it's causing a threading problem within Tapestry.
> Isn't the component creation multi-threaded?
>
>
> -Original Message-
> From: Howard Lewis Ship [mailto:hls...@gmail.com]
> Sent: Tuesday, December 23, 2008 2:00 PM
> To: Tapestry users
> Subject: Re: Tapestry Upgrade problem
>
> What JDK are you running on?
>
> On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman
> wrote:
>> I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6.
>> When I run my integration tests (using Selenium) I am get a error
> randomly.
>> I receive a org.apache.hivemind.ApplicationRuntimeException that
> indicates that given component (page) does not contain a component
> shell.
>> All of the affected pages have which leads me to
> believe that I have specified a shell component.
>> I have tracked the problem into the PageLoader class, but since the
> problem is random it is very difficult to debug.
>> Is it possible that this is a threading problem?
>>
>> Can anyone help?
>>
>> Mike
>>
>
>
>
> --
> Howard M. Lewis Ship
>
> Creator Apache Tapestry and Apache HiveMind
>
> -
> To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org
> For additional commands, e-mail: users-h...@tapestry.apache.org
>
>
> -
> To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org
> For additional commands, e-mail: users-h...@tapestry.apache.org
>
>
--
Andreas Andreou - andy...@apache.org - http://blog.andyhot.gr
<http://blog.andyhot.gr/> <http://blog.andyhot.gr/> <http://blog.andyhot.gr/>
Tapestry / Tacos developer
Open Source / JEE Consulting
-
To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org
For additional commands, e-mail: users-h...@tapestry.apache.org
-
To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org
For additional commands, e-mail: users-h...@tapestry.apache.org
RE: Tapestry Upgrade problem
I've tried a variety of JDKs and the problem happens with all of them. I've tracked it through the code and what it seems to come down to is that sometimes the _components Map in the AbstractComponent class is populated and other times it's null. My guess at this point is that it's some sort of threading problem within Tapestry. Anyone have any ideas on how to fix this problem? Thanks Mike From: Mike Wasserman [mailto:mike.wasser...@intelliware.ca] Sent: Thu 12/25/2008 1:07 PM To: Tapestry users Subject: RE: Tapestry Upgrade problem Thanks for your response. No, we didn't have this problem with 4.1.2 I am off on holidays for a little while, but I will try your suggestion as soon as I get back. Thanks again From: andre...@gmail.com on behalf of Andreas Andreou Sent: Wed 12/24/2008 3:11 PM To: Tapestry users Subject: Re: Tapestry Upgrade problem So, you weren't seeing this problem with 4.1.2 ? I'd be interesting if you could try with a 1.6 VM, but apart from that and since you're getting this consistently, i'm wondering if you can try this (which is based on some old reports I remember reading here): - Start your app and before running the tests, try to visit all the affected pages by hand (making sure they load fine) and only then fire the tests. Do you see problems after that? On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman wrote: > I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the > same problem with both. > The Selenium tests are hitting the app pretty hard and I'm wondering if > it's causing a threading problem within Tapestry. > Isn't the component creation multi-threaded? > > > -Original Message- > From: Howard Lewis Ship [mailto:hls...@gmail.com] > Sent: Tuesday, December 23, 2008 2:00 PM > To: Tapestry users > Subject: Re: Tapestry Upgrade problem > > What JDK are you running on? > > On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman > wrote: >> I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6. >> When I run my integration tests (using Selenium) I am get a error > randomly. >> I receive a org.apache.hivemind.ApplicationRuntimeException that > indicates that given component (page) does not contain a component > shell. >> All of the affected pages have which leads me to > believe that I have specified a shell component. >> I have tracked the problem into the PageLoader class, but since the > problem is random it is very difficult to debug. >> Is it possible that this is a threading problem? >> >> Can anyone help? >> >> Mike >> > > > > -- > Howard M. Lewis Ship > > Creator Apache Tapestry and Apache HiveMind > > - > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > > > - > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > > -- Andreas Andreou - andy...@apache.org - http://blog.andyhot.gr <http://blog.andyhot.gr/> <http://blog.andyhot.gr/> Tapestry / Tacos developer Open Source / JEE Consulting - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
RE: Tapestry Upgrade problem
Thanks for your response. No, we didn't have this problem with 4.1.2 I am off on holidays for a little while, but I will try your suggestion as soon as I get back. Thanks again From: andre...@gmail.com on behalf of Andreas Andreou Sent: Wed 12/24/2008 3:11 PM To: Tapestry users Subject: Re: Tapestry Upgrade problem So, you weren't seeing this problem with 4.1.2 ? I'd be interesting if you could try with a 1.6 VM, but apart from that and since you're getting this consistently, i'm wondering if you can try this (which is based on some old reports I remember reading here): - Start your app and before running the tests, try to visit all the affected pages by hand (making sure they load fine) and only then fire the tests. Do you see problems after that? On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman wrote: > I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the > same problem with both. > The Selenium tests are hitting the app pretty hard and I'm wondering if > it's causing a threading problem within Tapestry. > Isn't the component creation multi-threaded? > > > -Original Message- > From: Howard Lewis Ship [mailto:hls...@gmail.com] > Sent: Tuesday, December 23, 2008 2:00 PM > To: Tapestry users > Subject: Re: Tapestry Upgrade problem > > What JDK are you running on? > > On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman > wrote: >> I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6. >> When I run my integration tests (using Selenium) I am get a error > randomly. >> I receive a org.apache.hivemind.ApplicationRuntimeException that > indicates that given component (page) does not contain a component > shell. >> All of the affected pages have which leads me to > believe that I have specified a shell component. >> I have tracked the problem into the PageLoader class, but since the > problem is random it is very difficult to debug. >> Is it possible that this is a threading problem? >> >> Can anyone help? >> >> Mike >> > > > > -- > Howard M. Lewis Ship > > Creator Apache Tapestry and Apache HiveMind > > - > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > > > - > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > > -- Andreas Andreou - andy...@apache.org - http://blog.andyhot.gr <http://blog.andyhot.gr/> Tapestry / Tacos developer Open Source / JEE Consulting - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
Re: Tapestry Upgrade problem
So, you weren't seeing this problem with 4.1.2 ? I'd be interesting if you could try with a 1.6 VM, but apart from that and since you're getting this consistently, i'm wondering if you can try this (which is based on some old reports I remember reading here): - Start your app and before running the tests, try to visit all the affected pages by hand (making sure they load fine) and only then fire the tests. Do you see problems after that? On Wed, Dec 24, 2008 at 8:16 PM, Mike Wasserman wrote: > I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the > same problem with both. > The Selenium tests are hitting the app pretty hard and I'm wondering if > it's causing a threading problem within Tapestry. > Isn't the component creation multi-threaded? > > > -Original Message- > From: Howard Lewis Ship [mailto:hls...@gmail.com] > Sent: Tuesday, December 23, 2008 2:00 PM > To: Tapestry users > Subject: Re: Tapestry Upgrade problem > > What JDK are you running on? > > On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman > wrote: >> I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6. >> When I run my integration tests (using Selenium) I am get a error > randomly. >> I receive a org.apache.hivemind.ApplicationRuntimeException that > indicates that given component (page) does not contain a component > shell. >> All of the affected pages have which leads me to > believe that I have specified a shell component. >> I have tracked the problem into the PageLoader class, but since the > problem is random it is very difficult to debug. >> Is it possible that this is a threading problem? >> >> Can anyone help? >> >> Mike >> > > > > -- > Howard M. Lewis Ship > > Creator Apache Tapestry and Apache HiveMind > > - > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > > > - > To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org > For additional commands, e-mail: users-h...@tapestry.apache.org > > -- Andreas Andreou - andy...@apache.org - http://blog.andyhot.gr Tapestry / Tacos developer Open Source / JEE Consulting - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
RE: Tapestry Upgrade problem
I've tried with JDK 1.5.0_10 and JDK 1.5.0_17 and I am experiencing the same problem with both. The Selenium tests are hitting the app pretty hard and I'm wondering if it's causing a threading problem within Tapestry. Isn't the component creation multi-threaded? -Original Message- From: Howard Lewis Ship [mailto:hls...@gmail.com] Sent: Tuesday, December 23, 2008 2:00 PM To: Tapestry users Subject: Re: Tapestry Upgrade problem What JDK are you running on? On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman wrote: > I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6. > When I run my integration tests (using Selenium) I am get a error randomly. > I receive a org.apache.hivemind.ApplicationRuntimeException that indicates that given component (page) does not contain a component shell. > All of the affected pages have which leads me to believe that I have specified a shell component. > I have tracked the problem into the PageLoader class, but since the problem is random it is very difficult to debug. > Is it possible that this is a threading problem? > > Can anyone help? > > Mike > -- Howard M. Lewis Ship Creator Apache Tapestry and Apache HiveMind - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
RE: Tapestry Upgrade problem
java version "1.5.0_10" Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_10-b03) Java HotSpot(TM) Client VM (build 1.5.0_10-b03, mixed mode, sharing) -Original Message- From: Howard Lewis Ship [mailto:hls...@gmail.com] Sent: Tuesday, December 23, 2008 2:00 PM To: Tapestry users Subject: Re: Tapestry Upgrade problem What JDK are you running on? On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman wrote: > I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6. > When I run my integration tests (using Selenium) I am get a error randomly. > I receive a org.apache.hivemind.ApplicationRuntimeException that indicates that given component (page) does not contain a component shell. > All of the affected pages have which leads me to believe that I have specified a shell component. > I have tracked the problem into the PageLoader class, but since the problem is random it is very difficult to debug. > Is it possible that this is a threading problem? > > Can anyone help? > > Mike > -- Howard M. Lewis Ship Creator Apache Tapestry and Apache HiveMind - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
Re: Tapestry Upgrade problem
What JDK are you running on? On Tue, Dec 23, 2008 at 10:50 AM, Mike Wasserman wrote: > I am upgrading from Tapestry 4.1.2 to Tapestry 4.1.6. > When I run my integration tests (using Selenium) I am get a error randomly. > I receive a org.apache.hivemind.ApplicationRuntimeException that indicates > that given component (page) does not contain a component shell. > All of the affected pages have which leads me to believe > that I have specified a shell component. > I have tracked the problem into the PageLoader class, but since the problem > is random it is very difficult to debug. > Is it possible that this is a threading problem? > > Can anyone help? > > Mike > -- Howard M. Lewis Ship Creator Apache Tapestry and Apache HiveMind - To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
