Vincent Massol wrote:
> On Jan 11, 2008, at 10:02 PM, Sergiu Dumitriu wrote:
>
>> Vincent Massol wrote:
>>> On Jan 11, 2008, at 6:28 PM, sdumitriu (SVN) wrote:
>>>
>>>> Author: sdumitriu
>>>> Date: 2008-01-11 18:28:47 +0100 (Fri, 11 Jan 2008)
>>>> New Revision: 6782
>>>>
>>>> Modified:
>>>> xwiki-platform/core/trunk/xwiki-core/pom.xml
>>>> xwiki-platform/core/trunk/xwiki-core/src/main/java/com/xpn/xwiki/
>>>> render/groovy/GroovyTemplateEngine.java
>>>> xwiki-platform/core/trunk/xwiki-core/src/main/java/com/xpn/xwiki/
>>>> store/XWikiHibernateBaseStore.java
>>> [snip]
>>>
>>>> Modified: xwiki-platform/core/trunk/xwiki-core/src/main/java/com/
>>>> xpn/
>>>> xwiki/store/XWikiHibernateBaseStore.java
>>>> ===================================================================
>>>> --- xwiki-platform/core/trunk/xwiki-core/src/main/java/com/xpn/
>>>> xwiki/
>>>> store/XWikiHibernateBaseStore.java 2008-01-11 14:01:44 UTC (rev
>>>> 6781)
>>>> +++ xwiki-platform/core/trunk/xwiki-core/src/main/java/com/xpn/
>>>> xwiki/
>>>> store/XWikiHibernateBaseStore.java 2008-01-11 17:28:47 UTC (rev
>>>> 6782)
>>>> @@ -80,7 +80,7 @@
>>>> try {
>>>>
>>>> setHibUrl(XWiki.class.getClassLoader().getResource(path));
>>>> } catch (Exception ex2) {
>>>> - log.error("Failed setting the Hibernate
>>>> configuration file with any method, storage cannot be configured",
>>>> ex2);
>>>> + log.error("Failed setting the Hibernate
>>>> configuration file with any method, storage cannot be configured");
>>> Why are we removing the stack trace? This is an error and the stack
>>> trace will help debug the error.
>>>
>>> The fact that the test is wrongly written is a different matter IMO.
>>> The test must be forgetting to initialize something.
>>>
>>> But we do need the max amount of info when debugging an ERROR level
>>> error.
>>
>> It wasn't an error until a few days ago, when I made it so. It wasn't
>> even reported before, it was considered something that cannot happen
>> and
>> ignored.
>
> Sure but that doesn't answer the question. I see 2 solutions:
>
> 1) either it's not an error and shouldn't be reported as such (maybe
> warning, etc)
> 2) or it's an error and it should have a stack trace displayed.
>
> Said differently my rule is that we should always display stack traces
> for errors but we shouldn't display stack traces for warnings since
> warnings are acceptable whereas error are unexpected and means the
> application becomes unstable or it means that the user must do
> something to fix the problem since it's not a valid state.
>
> which is it here?
>
Well, it is an error, because Hibernate cannot be configured. And since
there are more possible causes (file doesn't exist, is not readable,
security restrictions forbid it, the configuration is wrong...), I
reckon a stacktrace would be useful.
Reverted.
Sergiu
_______________________________________________
devs mailing list
[email protected]
http://lists.xwiki.org/mailman/listinfo/devs
