Woa, that didn't work. According to a comment in the JSTL code, the Java developers don't know what their doing and the JSTL guys decided that the whole idea of a default resource bundle is non-portable. Not sure why, but that means that in order for this to work properly in both Struts and JSTL I have to maintain identical application.properties and application_en.properties files. That sounds like a major flaw in the JSTL design. Are their any of the JSTL designers around that would like to enlighten us on why this decision was made? (*Chris*)
On 11/9/06, Chris Pratt <[EMAIL PROTECTED]> wrote:
User evnafets over on the JSTL Forums at forum.java.sun.com tracked down the cause of the problem. In his words: The problem is JSTL is being helpful. Perhaps too helpful in this case. Basically it is looking for an exact match between ResourceBundle and your Locale settings. It will look for English first, but only find the default resource bundle. It will then look for French and find a specific resource bundle for French language. Thus JSTL maps itself to the french bundle because it is a more specific match (despite the fact the preferred language is English) If you have the JSTL source code, the class in question appears to be org.apache.taglibs.standard.tag.common.fmt.BundleSupport Specifically in the method findMatch(String basename, Locale pref) So I've added an empty application_en.properties file to my project and the problem seems to be closed. Thanks for everyone's help. (*Chris*) On 11/8/06, Chris Pratt <[EMAIL PROTECTED]> wrote: > > OK, now I'm scared. It's fixed and I can't figure out why? Earlier I > made the change to make sure that securityfilter came before I18NFilter in > the chain but the text was still showing up French. I put a bunch of > debugging code in (over hours and many, many tests) and now I can't > reproduce the problem, even after I've pulled out all my debugging code > (thinking something in there might have had an unanticipated side-effect). > Thanks for everybody's help and I'll definitely let everyone know if I > figure out what I did. > (*Chris*) > > On 11/8/06, Christopher Schultz < [EMAIL PROTECTED]> wrote: > > > > -----BEGIN PGP SIGNED MESSAGE----- > > Hash: SHA1 > > > > Paul, > > > > Paul Benedict wrote: > > > In Struts 1.3.6, I hope to be providing a locale solution for Struts > > > which is not tied to the session. > > > > Does struts rely on the session for locale information? I would have > > expected a null session to result in a call to request.getLocale for > > each request (or, at least, for each time the locale is required). > > > > I don't see why the session should be an absolute requirement. Is that > > > > currently the case? > > > > - -chris > > > > -----BEGIN PGP SIGNATURE----- > > Version: GnuPG v1.4.5 (MingW32) > > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org > > > > iD8DBQFFUnt89CaO5/Lv0PARAqqqAJ4ymycIXp1TH8p0r2zz4C+pTHQVkQCcCA3Q > > 7FLW3Y71MyxjvYqRD6xThLg= > > =vwCs > > -----END PGP SIGNATURE----- > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > >