[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-24 Thread richard.opa...@jboss.com
And I'd expect the following test to pass: --- public void testWebServicesDeployersOrder() throws Exception { DeployerClient main = createMainDeployer(); TestFlowDeployer deployer1 = new TestFlowDeployer("FakeDeployer"); deployer1.setOutputs("WebServicesMetaData", "WebServ

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-24 Thread richard.opa...@jboss.com
Cross reference: JBDEPLOY-201 View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4239758#4239758 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4239758 ___ jboss-user mailing list j

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"richard.opa...@jboss.com" wrote : | I'd expect this test to fail because there's no deployer with "1" output? Fail? The order is trivial - 1,2,3,4. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4239916#4239916 Reply to the post : http://www.jboss.org/inde

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"richard.opa...@jboss.com" wrote : And I'd expect the following test to pass: | Your test is wrong. #2 takes JBWMD as input, where #3,#4 declare that as output. So, this is properly ordered: 3,4,2 View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4239917#4239917

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"alesj" wrote : | #2 takes JBWMD as input, where #3,#4 declare that as output. | Ah, they all declare JBWMD as input and output. Then the name comparison kicks in - as the last resort on how to order. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4239919#

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread richard.opa...@jboss.com
"alesj" wrote : "alesj" wrote : | | Then the name comparison kicks in - as the last resort on how to order. | This is fundamentally wrong IMHO. | Current deployers ordering algorithm is broken. | You can't implement deployers ordering by just comparing two deployers. | Two deployers ar

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread richard.opa...@jboss.com
Current deployers sorting is broken because: - it doesn't take stages into account - it doesn't validate inputs - it sorts deployers on insert, but should sort them after deployers chain initialization - two deployers are not comparable in general View the original post : http://www.jboss.o

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread richard.opa...@jboss.com
"richard.opa...@jboss.com" wrote : | * sorting agorithm have to take stages, inputs, outputs and relative order into account | And deployers name comparison as last ordering decision ;) View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4239931#4239931 Repl

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread richard.opa...@jboss.com
Example of use case that shold fail but won't: Deployer1: - stage: PRE_REAL - inputs: output1 Deployer2: - stage: REAL - outputs: output1 PRE_REAL is before REAL. Current domino sorting algorithm will not detect it. View the original post : http://www.jboss.org/index.html?module=bb&op=viewt

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
What do you propose then? If we have to resort to name comparison it either means: * the order of how we process attachments doesn't matter * you are describing/ordering your deployers wrong == lack of info View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4239926

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"richard.opa...@jboss.com" wrote : | * deployer stages have to be comparable | Why? That's why they are called stages and they are already comparable - how else do you think they fit into our state machine. ;-) "richard.opa...@jboss.com" wrote : | * deployers have to be sorted on deplo

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"richard.opa...@jboss.com" wrote : | - it doesn't validate inputs | Validate how? Order at runtime based on the actual attachment values? Too much hassle for something that sounds useless in 99,9%. Not to mention that it's not deterministic. "richard.opa...@jboss.com" wrote : | - it so

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"richard.opa...@jboss.com" wrote : Example of use case that shold fail but won't: | | Deployer1: | - stage: PRE_REAL | - inputs: output1 | Deployer2: | - stage: REAL | - outputs: output1 | | PRE_REAL is before REAL. Current domino sorting algorithm will not detect it. I ne

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread alesj
"richard.opa...@jboss.com" wrote : Example of use case that shold fail but won't: | | Deployer1: | - stage: PRE_REAL | - inputs: output1 | Deployer2: | - stage: REAL | - outputs: output1 | | PRE_REAL is before REAL. Current domino sorting algorithm will not detect it. The

[jboss-user] [Microcontainer] - Re: Deployers Ordering

2009-06-25 Thread richard.opa...@jboss.com
In order to fix it the following is necessary to be done: * deployer stages have to be comparable * deployers have to be sorted on deployers chain retrieval (i.e. in method protected synchronized List getDeployersList(String stageName)) * sorting agorithm have to take stages, inputs, outputs