Re: svn commit: r475826 - /incubator/roller/trunk/build.xml

[Dave <sn...@gmail.com>]

On 11/16/06, agilliland@apache.org <ag...@apache.org> wrote:

>
>      <!-- Jar up web classes -->
> @@ -948,24 +948,21 @@
>          <jvmarg value="-Dro.build=${ro.build}"/>
>          <jvmarg value="-
> Droller.custom.config=${build.tests}/roller-custom.properties"/>
>          <formatter type="xml"/>
> -        <batchtest toDir="${build.reports}/business/xml">
> +        <batchtest toDir="${build.reports}/business/xml"
> haltonerror="true">



There are a couple of serious problems with haltonerror="true"

1) JUnit does not write the test report so there is no way to tell what went
wrong
2) build.xml does not stop the HSQLDB database, so the user must find and
kill the HSQLDB process

- Dave

Re: svn commit: r475826 - /incubator/roller/trunk/build.xml

[Allen Gilliland <Al...@Sun.COM>]

Dave wrote:
> On 11/16/06, agilliland@apache.org <ag...@apache.org> wrote:
> 
>>
>>      <!-- Jar up web classes -->
>> @@ -948,24 +948,21 @@
>>          <jvmarg value="-Dro.build=${ro.build}"/>
>>          <jvmarg value="-
>> Droller.custom.config=${build.tests}/roller-custom.properties"/>
>>          <formatter type="xml"/>
>> -        <batchtest toDir="${build.reports}/business/xml">
>> +        <batchtest toDir="${build.reports}/business/xml"
>> haltonerror="true">
> 
> 
> 
> There are a couple of serious problems with haltonerror="true"
> 
> 1) JUnit does not write the test report so there is no way to tell what 
> went
> wrong
> 2) build.xml does not stop the HSQLDB database, so the user must find and
> kill the HSQLDB process

hmmm, are you sure about #1?  when i was working on the business layer 
changes i had times when i got errors thrown from exceptions in the unit 
tests and it would properly log the errors in the report and stop the 
test.  and even if the error in the report isn't exactly what you were 
hoping for then then you should hopefully find something in the actual 
application logs under build/tests/log/roller.log

#2 is definitely a bummer, but there has to be a way to work around that 
in the build file right?  i would definitely think the starting/stopping 
of the HSQLDB would happen independently of the actual running of the tests.

i try looking into both of these.

-- Allen


> 
> - Dave
>