Provide alternative to the resource bundle mechanism

[schildbach]

Dnssecjava pulls strings from a messages.properties. Unfortunately, Android does not have this mechanism. This results in an exception whenever R.get() is called, e.g.:

Caused by: java.util.MissingResourceException: Can't find resource for bundle 'messages_en_US', key ''
at java.util.ResourceBundle.missingResourceException(ResourceBundle.java:238)
at java.util.ResourceBundle.getBundle(ResourceBundle.java:230)
at java.util.ResourceBundle.getBundle(ResourceBundle.java:139)
at org.jitsi.dnssec.R.<clinit>(R.java:20)

[schildbach]

@techguy613 You might want to have a look at this.

[ibauersachs]

Any suggestions on how you'd like to handle this instead?

[schildbach]

I think I would simply inline all the strings into the source code. If I understand it right, these are all exception messages which should not be translated anyway. So what's the resource bundle indirection good for?

Another idea would be a static setter that could be used by Android apps to set the resource bundle in your R class manually. Because Android actually can read resource bundles, it just doesn't support reading them from the classpath like you do.

[ibauersachs]

Hm, these were meant to be an explanation why a validation failed.

The setter is a good alternative, and as a fallback, returning the key (formatted like in the unit tests: key:param1:...:paramN)

[schildbach]

Ah, that fallback would be perfect. Even without the setter.

[techguy613]

Hey all, the setter sounds good to me. Should I take a crack at it or would you (@ibauersachs) like to make the change yourself?

[ibauersachs]

Some help would be very welcome :-)