You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@directory.apache.org by David Jencks <da...@yahoo.com> on 2007/08/23 09:50:01 UTC

possible test failures && a possible cure

I noticed on IRC while I was out elecharny  was getting some itest  
failures in mitosis with an error of

java.lang.IllegalStateException: No resource named org/apache/ 
directory/server/schema/bootstrap/partition/DBFILES

I don't know why I didn't see these the last time I made a change but  
they should generally be curable by adding this to the modules pom.xml:

     <dependency>
       <groupId>org.apache.directory.server</groupId>
       <artifactId>apacheds-bootstrap-partition</artifactId>
       <version>${pom.version}</version>
       <scope>test</scope>
     </dependency>

Please don't leave out the <scope>test</scope> as that will make it a  
lot harder to swap bootstrap partitions.

I fixed the problems in mitosis.  If anyone runs into  similar  
problems feel free to ping me on irc, if I'm around I'm glad to help.

I still got a test failure earlier in the build in   
org.apache.directory.server.BadDnTest.

thanks
david jencks


Re: possible test failures && a possible cu

Posted by Alex Karasulu <ak...@apache.org>.
On 8/23/07, David Jencks <da...@yahoo.com> wrote:
>
>
> On Aug 23, 2007, at 6:09 AM, Alex Karasulu wrote:
>
> David this change you made on the bootstrap-partition dependency in the
> core from
> compile scope to "test" scope will have greater ramifications as I noted
> in a few other
> emails.
>
> You'll need to fix server-installers, and server-main for one and possibly
> other projects
> which presume that this dependency is transitively inherited.
>
>
> I'm confused about this.  In the past I've asked how to tell if a change I
> propose breaks anything and the response has always been, "run the
> integration tests, if they pass nothing is broken".  Well, I ran all the
> integration tests before committing this change, and nothing broke.  If
> there's something else I can do to determine if the server is working,
> please let me know what it is.  Otherwise, I have to say I can't understand
> why you think any other changes are needed.
>

Just for a conclusion (we figured this out on IRC) let me comment:

I we got to the bottom of this on IRC.  The reason why you missed this hairy
little trap
even with the integration test runs was because the server-installers and
server-main
modules are not included in the build with the default maven profile.

I guess you're going to change this by making the default profile run all
integ tests and
include all the modules.  Then we can trigger just a quick build for
developers to quickly
compile modules.

Good idea!!! We should have done this from the start.

Alex

Re: possible test failures && a possible cur

Posted by David Jencks <da...@yahoo.com>.
On Aug 23, 2007, at 6:09 AM, Alex Karasulu wrote:

> David this change you made on the bootstrap-partition dependency in  
> the core from
> compile scope to "test" scope will have greater ramifications as I  
> noted in a few other
> emails.
>
> You'll need to fix server-installers, and server-main for one and  
> possibly other projects
> which presume that this dependency is transitively inherited.

I'm confused about this.  In the past I've asked how to tell if a  
change I propose breaks anything and the response has always been,  
"run the integration tests, if they pass nothing is broken".  Well, I  
ran all the integration tests before committing this change, and  
nothing broke.  If there's something else I can do to determine if  
the server is working, please let me know what it is.  Otherwise, I  
have to say I can't understand why you think any other changes are  
needed.

The main idea behind separating bootstrap-partition from bootstrap- 
extract was that a project could specify exactly which bootstrap  
partition it wanted to use, so projects should probably not be  
assuming that their bootstrap partition is inherited "for free"

Maybe the problem is that there is no test that sees if the server  
that's going to be packed up into the installer actually starts?

>
> Also it would be nice to throw a nice descriptive exception that a  
> bootstrap partition jar is not present instead of complaining that  
> the DBFILES are missing.  This will prevent much noise from our  
> users.  Could you look into adding something like this before we  
> release?

Great idea, see rev. 569056

thanks
david jencks

>
> Thanks,
> Alex
>
> On 8/23/07, David Jencks <da...@yahoo.com> wrote:
> I noticed on IRC while I was out elecharny  was getting some itest
> failures in mitosis with an error of
>
> java.lang.IllegalStateException: No resource named org/apache/
> directory/server/schema/bootstrap/partition/DBFILES
>
> I don't know why I didn't see these the last time I made a change but
> they should generally be curable by adding this to the modules  
> pom.xml:
>
>      <dependency>
>        <groupId>org.apache.directory.server </groupId>
>        <artifactId>apacheds-bootstrap-partition</artifactId>
>        <version>${pom.version}</version>
>        <scope>test</scope>
>      </dependency>
>
> Please don't leave out the <scope>test</scope> as that will make it a
> lot harder to swap bootstrap partitions.
>
> I fixed the problems in mitosis.  If anyone runs into  similar
> problems feel free to ping me on irc, if I'm around I'm glad to help.
>
> I still got a test failure earlier in the build in
> org.apache.directory.server.BadDnTest.
>
> thanks
> david jencks
>
>


Re: possible test failures && a possible cur

Posted by Alex Karasulu <ak...@apache.org>.
David this change you made on the bootstrap-partition dependency in the core
from
compile scope to "test" scope will have greater ramifications as I noted in
a few other
emails.

You'll need to fix server-installers, and server-main for one and possibly
other projects
which presume that this dependency is transitively inherited.

Also it would be nice to throw a nice descriptive exception that a bootstrap
partition jar is not present instead of complaining that the DBFILES are
missing.  This will prevent much noise from our users.  Could you look into
adding something like this before we release?

Thanks,
Alex

On 8/23/07, David Jencks <da...@yahoo.com> wrote:
>
> I noticed on IRC while I was out elecharny  was getting some itest
> failures in mitosis with an error of
>
> java.lang.IllegalStateException: No resource named org/apache/
> directory/server/schema/bootstrap/partition/DBFILES
>
> I don't know why I didn't see these the last time I made a change but
> they should generally be curable by adding this to the modules pom.xml:
>
>      <dependency>
>        <groupId>org.apache.directory.server</groupId>
>        <artifactId>apacheds-bootstrap-partition</artifactId>
>        <version>${pom.version}</version>
>        <scope>test</scope>
>      </dependency>
>
> Please don't leave out the <scope>test</scope> as that will make it a
> lot harder to swap bootstrap partitions.
>
> I fixed the problems in mitosis.  If anyone runs into  similar
> problems feel free to ping me on irc, if I'm around I'm glad to help.
>
> I still got a test failure earlier in the build in
> org.apache.directory.server.BadDnTest.
>
> thanks
> david jencks
>
>