Like I thought I didn't got the exact point. Thx for clarification. I'm totally with you. (but I still don't like the guide exceptions - it is hard to write a highly customisable application around it) On Mar 30, 2013, at 10:51 AM, Mikhail Mazursky <mikhail.mazur...@gmail.com> wrote:
> Hi, > > advantages are Guice shows: > - where the offending constant was bound; > - to what type it tried to convert it (not necessary the type that is > required - can be subtype); > - where the converter was bound. > > I think this can be helpfull information :) > > 2013/3/30 Daniel Manzke <daniel.man...@googlemail.com> > >> Hi, >> >> I don't see any advantage from 2. But I'm not familiar with the test case >> maybe I just don't get the point of the information which is not delivered >> with 1 but in 2. >> Could you explain it a little bit more? :) >> >> >> Thx, >> Daniel >> >> On Mar 30, 2013, at 9:45 AM, Mikhail Mazursky (JIRA) <j...@apache.org> >> wrote: >> >>> >>> [ >> https://issues.apache.org/jira/browse/ONAMI-88?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13618021#comment-13618021] >>> >>> Mikhail Mazursky commented on ONAMI-88: >>> --------------------------------------- >>> >>> I experimented a bit with catching RuntimeException and wrapping it in >> ProvisionException vs just letting exception propagate. I used existing >> code in UUIDConverter as a testcase. So, the current behaviour (catch&wrap) >> results in this output (i obused a test in other project for experiments - >> don't take it into account): >>> >>> com.google.inject.ConfigurationException: Guice configuration errors: >>> >>> 1) String value 'abcasd' is not a valid UUID >>> while locating java.util.UUID annotated with >> @com.google.inject.name.Named(value=a) >>> for field at >> org.apache.onami.converters.net.URLConverterTestCase.u(URLConverterTestCase.java:43) >>> while locating org.apache.onami.converters.net.URLConverterTestCase >>> >>> 1 error >>> at >> com.google.inject.internal.InjectorImpl.getProvider(InjectorImpl.java:1004) >>> [...] >>> Caused by: java.lang.IllegalArgumentException: Invalid UUID string: >> abcasd >>> at java.util.UUID.fromString(UUID.java:194) >>> [...] >>> >>> If we don't catch - this is the output: >>> >>> com.google.inject.ConfigurationException: Guice configuration errors: >>> >>> 1) Error converting 'abcasd' (bound at >> org.apache.onami.converters.net.URLConverterTestCase$1.configure(URLConverterTestCase.java:54)) >> to java.util.UUID >>> using java.util.UUID which matches only(java.util.UUID) (bound at >> org.apache.onami.converters.core.AbstractConverter.configure(AbstractConverter.java:44)). >>> Reason: java.lang.IllegalArgumentException: Invalid UUID string: abcasd >>> while locating java.util.UUID annotated with >> @com.google.inject.name.Named(value=a) >>> for field at >> org.apache.onami.converters.net.URLConverterTestCase.u(URLConverterTestCase.java:43) >>> while locating org.apache.onami.converters.net.URLConverterTestCase >>> >>> 1 error >>> at >> com.google.inject.internal.InjectorImpl.getProvider(InjectorImpl.java:1004) >>> [...] >>> Caused by: java.lang.IllegalArgumentException: Invalid UUID string: >> abcasd >>> at java.util.UUID.fromString(UUID.java:194) >>> [...] >>> >>> As you can see, wrapping in ProvisionException gives us less information >> - apparently Guice treats ProvisionException somehow not like other >> exceptions. I propose to get rid of wrapping in ProvisionException and wrap >> only checked exceptions in RuntimeException where necessary. WDYT? >>> >>>> Consistent policy for exception handling >>>> ---------------------------------------- >>>> >>>> Key: ONAMI-88 >>>> URL: https://issues.apache.org/jira/browse/ONAMI-88 >>>> Project: Apache Onami >>>> Issue Type: Improvement >>>> Components: configuration >>>> Affects Versions: configuration-6.3.0 >>>> Reporter: Mikhail Mazursky >>>> Priority: Minor >>>> >>>> No consistent policy for exception handling - for example >> SQLTimestampConverter catches Throwable and rethrows it as >> ProvisionException but CharsetConverter do not catch anything at all. >> Should we catch at all or Guice handles this itself? >>>> I guess we should rethrow only checked exceptions. Also no need to >> catch Errors in such places in any case (catch Exception vs Throwable). >>> >>> -- >>> This message is automatically generated by JIRA. >>> If you think it was sent incorrectly, please contact your JIRA >> administrators >>> For more information on JIRA, see: >> http://www.atlassian.com/software/jira >> >>
