struts-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christian Balzer <>
Subject org.apache.struts2.interceptor.I18nInterceptor seems to work only once + can defaulting to accept headers (and impacting on number/date parsing) be disabled?
Date Wed, 12 Jul 2017 23:01:20 GMT
Hi all,

I joined a project recently that makes use of Struts 2, and was tasked
with enabling internationalisation for it. I'm completely new to the
framework, and the project has organically grown over the years -
which means the project has a multitude of interceptors and chains
configured (some interceptors even seem to be called more than once by
the same chain), which might interfere with each other. So please bear
with me.

After spending some time bringing the project to the latest version of
Struts (and dealing with several breaking changes, mostly on the
Struts tags level), I have added the I18nInterceptor, as described on
[1] and added the following two parameters to it:

Our resource bundle has an English default and a "de" properties file.

I also set ParametersInterceptor's excludeParams to "language" as per
[2], to avoid getting errors about my actions not having a "language"
setter. (For some reason, that appeared in the log, where I would have
expected it, albeit not as an ERROR, and on the page itself, which
prompted me to disable it.) Was that the right thing to do, or

When I set my "default" browser language to German, and call the page,
it loads the German text all right (so L10n of text works).
When I append the page's URL by ?language=en, it switches back to
English, which is great.
Sadly. this only seems to work exactly once per session. When I then
try to call ?language=de again, it stays English.
That's problem number one.

Problem number two is that I would prefer the interceptor to not check
the accept header, i.e. to ignore the browser.
Part of the reason is that the project owners always want to display
English, unless another language is explicitly chosen via URL
parameter. But mainly, we use number and date parsing in a lot of
places, and that depends on a British locale (i.e. en_GB, or de_GB;
the JS front-end doesn't check the user's locale, and calendar widgets
etc. all use British formats). A colleague of mine accessed the test
server after we had enabled the interceptor, and his browser had en_US
set as the first language in his accept header. Let's just say it
didn't go well, because our app didn't know what to make of the 12th
day of the 31st month anno 2017... (Ideally, we'd go for full g11n,
i.e. i18n of everything, not just text messages, but we won't have
time for it for the foreseeable future. Instead, for the immediate
future, we'd like maybe to extend the interceptor to only accept
locales from a predefined list (forcing the country to GB) - or only
use the locale for getText() [3] and nothing else...) Is there a
(fairly straightforward) way to disable the accept header fallback?

Last but not least, I have been unable to get debug output from either
of the two interceptors mentioned [4]...

I'd really appreciate if anyone could nudge me into the right
direction - especially regarding the "works once, but only once"

Kind regards,

P.S.: If all else fails, I thought I might try to use a custom-made
language switch similar to [5]...?


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message