You are viewing a plain text version of this content. The canonical link for it is here.
Posted to slide-dev@jakarta.apache.org by Oliver Zeigermann <oz...@c1-fse.de> on 2004/06/21 14:17:51 UTC
Slide initialization broken?
With the latest version from the CVS Slide no longer is able to
initialize itself (at least with caching turned off). This worked with
milestone 2.1M1 and thus this must have been caused by commits done this
month.
What I get is an exception caught in XMLMarshaller saying "Insufficient
credentials to create object".
Inspecting recent commits I could not find anything that could have
causing this problem. So, recent committers:
- Stefan (Lützkendorf)
- Peter
- Jürgen
could you please check?
Thanks,
Oliver
---------------------------------------------------------------------
To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: slide-dev-help@jakarta.apache.org
What are SlideTokenWrappers good for in the first place? WAS Severe
Problem with authentication code / WAS Slide initialization broken?
Posted by Oliver Zeigermann <oz...@c1-fse.de>.
I globally disabled disabling of force enlistment in the
SlideTokenWrapper and now everything works fine (at least after I fixed
another flaw in ExtendedStore).
This makes me wonder, what is this wrapping and changing of force
enlistment good for in the first place? Can I check in the disabled
version or does this cause any trouble?
Olliver
Oliver Zeigermann wrote:
> Ok, the code I had added was to fix a flaw in TxXMLFileDescriptorsStore.
> When a transaction gets suspended from a thread the active transaction
> branch has to be cleared by
>
>> activeTransactionBranch.set(null);
>
>
> And that's what I did. After that I got the access denied excpetion with
> the "Insufficient credentials to create object" message. How can this
> be? The problem is the code in the checkCredentials method in class
> SecurityImpl. The Slide token is set to read only with
>
>> else if (token.isForceStoreEnlistment()) {
>> token = new SlideTokenWrapper(token, false); // read only
>> access
>> }
>
>
> and the retrieve request is thus outside of any transaction. However,
> upon initial import all data is created inside one big transaction, but
> security checking is done outside of it simultaneously. Of course
> checking of data not yet committed outside of the transaction the data
> has been created must fail, which is what happens with my code. Still,
> however, without the above flaw I was trying to fix, the checking still
> *does* happen in the same transaction and that's why it had been working
> before.
>
> Now, my question is: Why is the checking of credentials done in read
> only mode anyway? Is it save to simply remove it?
>
> Cheers,
>
> Oliver
>
>
> Oliver Zeigermann wrote:
>
>> Hehehehe, seems I have forgotten one person: myself ;)
>>
>> This was caused by code I had modified locally :(
>>
>> So, sorry for the noise and forget about it...
>>
>> Oliver
>>
>> Oliver Zeigermann wrote:
>>
>>> With the latest version from the CVS Slide no longer is able to
>>> initialize itself (at least with caching turned off). This worked
>>> with milestone 2.1M1 and thus this must have been caused by commits
>>> done this month.
>>>
>>> What I get is an exception caught in XMLMarshaller saying
>>> "Insufficient credentials to create object".
>>>
>>> Inspecting recent commits I could not find anything that could have
>>> causing this problem. So, recent committers:
>>>
>>> - Stefan (Lützkendorf)
>>> - Peter
>>> - Jürgen
>>>
>>> could you please check?
>>>
>>> Thanks,
>>>
>>> Oliver
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
>>> For additional commands, e-mail: slide-dev-help@jakarta.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
>> For additional commands, e-mail: slide-dev-help@jakarta.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: slide-dev-help@jakarta.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: slide-dev-help@jakarta.apache.org
Severe Problem with authentication code / WAS Slide initialization
broken?
Posted by Oliver Zeigermann <oz...@c1-fse.de>.
Ok, the code I had added was to fix a flaw in TxXMLFileDescriptorsStore.
When a transaction gets suspended from a thread the active transaction
branch has to be cleared by
> activeTransactionBranch.set(null);
And that's what I did. After that I got the access denied excpetion with
the "Insufficient credentials to create object" message. How can this
be? The problem is the code in the checkCredentials method in class
SecurityImpl. The Slide token is set to read only with
> else if (token.isForceStoreEnlistment()) {
> token = new SlideTokenWrapper(token, false); // read only access
> }
and the retrieve request is thus outside of any transaction. However,
upon initial import all data is created inside one big transaction, but
security checking is done outside of it simultaneously. Of course
checking of data not yet committed outside of the transaction the data
has been created must fail, which is what happens with my code. Still,
however, without the above flaw I was trying to fix, the checking still
*does* happen in the same transaction and that's why it had been working
before.
Now, my question is: Why is the checking of credentials done in read
only mode anyway? Is it save to simply remove it?
Cheers,
Oliver
Oliver Zeigermann wrote:
> Hehehehe, seems I have forgotten one person: myself ;)
>
> This was caused by code I had modified locally :(
>
> So, sorry for the noise and forget about it...
>
> Oliver
>
> Oliver Zeigermann wrote:
>
>> With the latest version from the CVS Slide no longer is able to
>> initialize itself (at least with caching turned off). This worked with
>> milestone 2.1M1 and thus this must have been caused by commits done
>> this month.
>>
>> What I get is an exception caught in XMLMarshaller saying
>> "Insufficient credentials to create object".
>>
>> Inspecting recent commits I could not find anything that could have
>> causing this problem. So, recent committers:
>>
>> - Stefan (Lützkendorf)
>> - Peter
>> - Jürgen
>>
>> could you please check?
>>
>> Thanks,
>>
>> Oliver
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
>> For additional commands, e-mail: slide-dev-help@jakarta.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: slide-dev-help@jakarta.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: slide-dev-help@jakarta.apache.org
Re: Slide initialization broken?
Posted by Oliver Zeigermann <oz...@c1-fse.de>.
Hehehehe, seems I have forgotten one person: myself ;)
This was caused by code I had modified locally :(
So, sorry for the noise and forget about it...
Oliver
Oliver Zeigermann wrote:
> With the latest version from the CVS Slide no longer is able to
> initialize itself (at least with caching turned off). This worked with
> milestone 2.1M1 and thus this must have been caused by commits done this
> month.
>
> What I get is an exception caught in XMLMarshaller saying "Insufficient
> credentials to create object".
>
> Inspecting recent commits I could not find anything that could have
> causing this problem. So, recent committers:
>
> - Stefan (Lützkendorf)
> - Peter
> - Jürgen
>
> could you please check?
>
> Thanks,
>
> Oliver
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: slide-dev-help@jakarta.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: slide-dev-help@jakarta.apache.org