Thanks for sharing your thoughts! I reflected on this issue again and I think that the excludeFields array is the root of my problem. Now you can only tell reflectionEquals which fields to exclude, but you never know which non-transient, non-static, non-$ fields are included at runtime. That makes the result of reflectionEquals unpredictable. The only way to guarantee a consistent result would be to use an includeFields array instead.
I have to release tomorrow, so for now I decided to replace all occurrences of reflectionEquals with the good old copy-paste logic. I like the simplicity and elegance of reflectionEquals, so I hope that we can work this out. Cheers, Bas -----Oorspronkelijk bericht----- Van: [email protected] [mailto:[email protected]] Namens Paul Benedict Verzonden: woensdag 27 juli 2011 20:52 Aan: Commons Users List Onderwerp: Re: [lang] Advise needed for broken reflectionEquals in WebLogic 10.3.5 +1 to that idea. On Wed, Jul 27, 2011 at 12:25 PM, Gary Gregory <[email protected]> wrote: > We might as well use an RE instead of a prefix. Next week someone will want > to do something will fields that /end/ in "_" or whatnot. > > Gary > > -----Original Message----- > From: [email protected] [mailto:[email protected]] > On Behalf Of Paul Benedict > Sent: Wednesday, July 27, 2011 13:15 PM > To: Commons Users List > Subject: Re: [lang] Advise needed for broken reflectionEquals in > WebLogic 10.3.5 > > ReflectionEquals uses actual fields, not methods, to build the string output. > The underscore is actually a known convention by some Java programmers (I > don't subscribe to it but projects like Apache Tapestry use underscore for > ALL their fields). So I don't think underscore is suitable to be excluded. > However, perhaps as an enhancement the builder could be configured with a > list of prefixes to ignore. If that's what you want, please open up a JIRA > issue for consideration. > > On Wed, Jul 27, 2011 at 11:49 AM, Cancrinus, Bas <[email protected]> > wrote: >> I'm currently migrating an EAR from WebLogic 10.3.3 to 10.3.5. While testing >> I noticed that the following invocation returned false while I expected true >> (lang 2.5): >> >> EqualsBuilder.reflectionEquals(Object lhs, Object rhs, String[] >> excludeFields) >> >> I noticed in my debugger that my domain classes were extended by the >> WebLogic JPA provider with the following field: >> >> protected org.eclipse.persistence.queries.FetchGroup >> _persistence_fetchGroup >> >> EqualsBuilder.reflectionAppend() (lang 2.5-3.0) only ignores: >> >> - "$" in field names >> >> - static modifiers >> >> - transients (conditional) >> >> - and of course the excludeFields array >> >> The JPA extension field mentioned above doesn't fall into any of those >> categories, so it is appended to the equals result. >> >> The easy way out for me would be to add this field to all reflectionEquals >> excludeFields arrays, but I'm looking for a more sustainable way to do this. >> >> - What do you think about patching EqualsBuilder.reflectionAppend() >> so that it ignores this field ( f.startsWith("_") )? >> >> - Are there any other better ways to fix this issue? >> >> I appreciate any help very much! >> >> Have a nice day, >> Bas >> >> >> P: Please consider the environment before printing this e-mail >> >> ________________________________ >> >> This e-mail may contain confidential and privileged material. You are >> requested not to disclose, copy or distribute any information thereof. If >> you are not the intended recipient (or have received this e-mail in error) >> please notify the sender immediately and delete this e-mail. We accept no >> liability for damage related to data and/or documents which are communicated >> by electronic mail. >> >> ________________________________ >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] P: Please consider the environment before printing this e-mail ________________________________ This e-mail may contain confidential and privileged material. You are requested not to disclose, copy or distribute any information thereof. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. We accept no liability for damage related to data and/or documents which are communicated by electronic mail. ________________________________ --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
