[
https://issues.jboss.org/browse/DROOLS-1039?page=com.atlassian.jira.plugi...
]
Geoffrey De Smet edited comment on DROOLS-1039 at 4/14/16 6:56 AM:
-------------------------------------------------------------------
I think it should fail fast, so it should throw an Exception, such as an
IllegalStateException.
rantStart
Log messages can get lost (especially when deploying on app servers that ignore your
logging config), are frequently ignored (too many projects start up with a dozen log
messages, so that log message will scroll by in half a second, ...).
Crashing the app is a much better solution. It forces the developer to fix his stuff,
immediately after he broke it, while he/she still remembers what he/she was doing. It
doesn't allow it result to be judged as "assemblers don't work" (when
the error message is buried in the log).
Frameworks that don't fail fast, always have the same destiny in large enterprise
projects: they gather misconfigurations over time and their results/behaviour become more
and more surprising (= incorrect) due to these misconfigurations. Users lose faith in them
and even though the users are to blame for the misconfigurations, I believe the framework
developers are too blame for allowing them to get away with it :)
rantEnd
was (Author: ge0ffrey):
I think it should fail fast, so it should throw an Exception, such as an
IllegalStateException.
rantStart
Log messages can get lost (especially when deploying on app servers that ignore your
logging config), are frequently ignored (too many projects start up with a dozen log
messages, so that log message will scroll by in half a second, ...).
Crashing the app is a much better solution. It forces the developer to fix his stuff,
immediately after he broke it, while he/she still remembers what he/she was doing. It
doesn't allow it result to be judged as "assemblers don't work" (when
the error message is buried in the log).
Frameworks that don't fail fast, always have the same destiny in large enterprise
projects: they gather misconfigurations over time and they results/behaviour becomes
surprising (as in incorrect) due to this misconfigurations. Users lose faith in them and
even those the users are to blame for the misconfigurations, I believe the framework
developers are too blame for allowing them to get away with it :)
rantEnd
An invalid kie.conf should throw an exception and not be silently
ignored
-------------------------------------------------------------------------
Key: DROOLS-1039
URL:
https://issues.jboss.org/browse/DROOLS-1039
Project: Drools
Issue Type: Enhancement
Components: core engine
Affects Versions: 6.4.0.Beta1
Reporter: Geoffrey De Smet
Assignee: Mario Fusco
In drools-belief or optaplanner-core, change the META-INF/kie.conf file like this:
{code}
[
'assemblers' : [ new org.foo.ThisDoesNotExists() ],
'weavers' : [],
'runtimes' : [],
'beliefs' : []
]
{code}
Current behavior: it doesn't throw any exception, but of course later on things go
haywire at some point with a completely unrelated error message.
Expected Behavior: Fail fast and let me know that org.foo.ThisDoesNotExists does not
exist and that my kie.conf is in error.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)