[Chandler-dev] Localization loading discussion

[Phillip J. Eby]

Just as an FYI, the prototype that Mikku has been working on is a 
demonstration of something similar to my original proposal for this.  That 
is, an API that can find all the egg-supplied resources at startup with 
only a single scan.  It doesn't need the repository or any special install 
steps.  So I'd suggest looking at his prototype to decide where to go next.

At 02:09 PM 6/7/2006 -1000, Brian Kirsch wrote:
>Hello,
>There is an ongoing discussion regarding bug 6007 "Need to localize code 
>that is run before repository is up".
>
>
><http://bugzilla.osafoundation.org/show_bug.cgi?id=6007>http://bugzilla.osafoundation.org/show_bug.cgi?id=6007
>
>Since others may have thoughts and opinions on the subject I am forwarding 
>to the list.
>
>Basically it comes down to the issue that Chandler needs access to certain 
>localized strings and resources
>before the Repository is created.
>
>
>
>------- Comment #5 >From <mailto:bkirsch at osafoundation.org>Brian Kirsch 
>2006-06-07 17:07 PST 
>[<https://bugzilla.osafoundation.org/show_bug.cgi?id=6007#add_comment>reply] 
>-------
>Yes Markku is correct,
>We are *not* storing translations in the repository.
>
>There was discussion of storing the paths to translation files in the repo and
>registering translations at installParcel time. But frankly, I am not conviced
>this is the right approach at least for the Chandler level translations since
>that still leaves a large edge case where we need translation strings before
>the repo is created.
>
>I am leaning at this point to just having a know directory structure for
>Chandler level translations and resources and using scripts to install
>localizations in to that structure.
>
>Third party devs could use the installParcel mechanism since the repo will be
>present at that point.
>
>An alternate approach is to have a registration file that lives outside of the
>repo and contains the locations of Chandler level translations and resources.
>That was a developer could add a translation for Chandler outside the Chandler
>directory structure and just append the location to the file.
>
>I think we should avoid the repo at all costs for Chandler level translations
>as using the repo implies that we have to have another mechanism for loading
>pre-repo strings and images.
>
>There is nothing wrong with an item /block storing its translations in the 
>repo
>however.
>
>As long as we have a means to refresh the data if the locale or schema version
>changes.
>
>
>
>
>
>------- Comment #4 >From <mailto:mmmm at osafoundation.org>Markku Mielityinen 
>2006-06-07 14:50 PST 
>[<https://bugzilla.osafoundation.org/show_bug.cgi?id=6007#add_comment>reply] 
>-------
>Do you see some benefit in this approach? Items that are stored to repo 
>use _()
>when they are created. Obviously these items work only when repo is up. Items
>that are not stored in the repo can/and should use _() each time they are
>created (that is each time that you run the app at least). And _() service is
>up all the time. I don't see a translation case when you would like to
>store/load to/from repo when it is not up; obviously one exists but I just
>cannot see it. In my eyes repo and translations (and other resources) are more
>or less orthogonal. The fact that we have translated string in repo is more an
>effect than cause (causality). I don't see a reason to put all translations to
>repo. Maybe you want a CPIA style splashscreen or something like that but 
>to me
>this seems more or less secondary. So an example please :)
>
>
>
>------- Comment #3 >From <mailto:heikki at osafoundation.org>Heikki Toivonen 
>2006-06-07 14:24 PST 
>[<https://bugzilla.osafoundation.org/show_bug.cgi?id=6007#add_comment>reply] 
>-------
>The problem is that we intend to store the translations in the repository. 
>This
>means the boostrap code has no access to translations since we don't have a
>repository yet. So we need a different way of translating the bootstrap code.
>
>------- Comment #2 >From <mailto:mmmm at osafoundation.org>Markku Mielityinen 
>2006-06-07 14:17 PST 
>[<https://bugzilla.osafoundation.org/show_bug.cgi?id=6007#add_comment>reply] 
>-------
>I am not sure I that I understand the problem. _() translation macro should
>work right from the start, right? At least mine does and it loads translations
>from an egg according to the locale setting. And all resources work the same
>way. I think that you can even have localized XRC based dialogs if you want to
>:) Please give a more detailed description of the problem. This questions
>relates to <https://bugzilla.osafoundation.org/show_bug.cgi?id=4656>bug 
>4656 as well.
>
>
>
>------- Comment #1 >From <mailto:bkirsch at osafoundation.org>Brian Kirsch 
>2006-06-07 13:31 PST 
>[<https://bugzilla.osafoundation.org/show_bug.cgi?id=6007#add_comment>reply] 
>-------
>yes this is a known issue which I am currently thinking about how best to 
>solve
>from an architectual standpoint.
>
>
>
>
>
>
>
>
>
>
>
>--
>Brian Kirsch -  Cosmo Developer / Chandler Internationalization Engineer
>Open Source Applications Foundation
>543 Howard St. 5th Floor
>San Francisco, CA 94105
><http://www.osafoundation.org>http://www.osafoundation.org
>
>_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>
>Open Source Applications Foundation "chandler-dev" mailing 
>listhttp://lists.osafoundation.org/mailman/listinfo/chandler-dev