On Tue, Nov 24, 2009 at 2:10 PM, Thadeus Burgess <thade...@thadeusb.com>wrote:
> Why can't it just be 'en'?
>
It probably could.... I just checked the request environment in a "new
install" browser I've never used ("Konquerer on Ubuntu) and web2py is
picking up
http_accept_language=['en-US', 'en']
So if a client had ['en-UK', 'en'], if no translation file for en-uk.py
existed, it would "pick up" the en, and deliver in the site / apps
internally encoded strings (which would be appropriate).
If this is consistent, that the ordering from a client is first
country-specific, then country-agnostic, then this would probably be
reasonable. The only downside: if I complained (from UK for example)
about "color" being misspelled (if I think it should be "colour"), then the
app is not being explicit enough about what it says it's servering.
Having said that, I am now convinced (pretty well) that 'en' should not be
part of the gluon/languages initialization. I think it should be explicit,
and only one language - the more specific declaration, not the broader one.
So - I think that the original patch I sent (with only 'en-us') is correct,
and what we should use.
- Yarko
> -Thadeus
>
>
>
>
>
> On Tue, Nov 24, 2009 at 1:01 PM, Yarko Tymciurak <
> resultsinsoftw...@gmail.com> wrote:
>
>> a short version of this (context: U.N. type of meeting):
>>
>> you (in effect) changed the translator initialization code to say "The
>> default language [string] I will present you is TWO lanugages", Massimo, it
>> is _as if_ you said something like: "I'm speaking Itailan, or another way
>> for you to think of it - I am speaking Russian"
>>
>> It cannot be! I cannot "hear" you that way - I have to know _which_
>> language, if I am to have any hope of "hiring' the right translator!
>>
>> There is no "can of worms" in the _problem domain_; it is in your not
>> being specific enough in what you told me you would be speaking!
>>
>> :-)
>>
>>
>>
>>
>> On Tue, Nov 24, 2009 at 12:51 PM, Yarko Tymciurak <
>> resultsinsoftw...@gmail.com> wrote:
>>
>>>
>>> On Tue, Nov 24, 2009 at 12:37 PM, mdipierro <mdipie...@cs.depaul.edu>wrote:
>>>
>>>>
>>>> I put en-uk not en-gr.
>>>
>>>
>>> right, en-uk; from a software analysis perspective, only one language
>>> should be in the initialization (addition of a non-country specific version
>>> of said language should also be acceptable).
>>>
>>> You see, this is a big can of worms. How do you
>>>> know that the default application is in en-us and not en-uk?
>>>
>>>
>>> This is not can of worms at all: you do not "know" - you declare; the
>>> patch as you made it, you in effect declared TWO languages (two contry
>>> specific versions of the same base language, but for understanding this it
>>> is clearer to ignore the "non-country-specific" part - and just think of it
>>> as TWO languages.
>>>
>>> When you look at this as TWO languages, and your translation class code,
>>> you will see that once any language is in accepted languages, it will not be
>>> picked up from the application's languages/*.py file.
>>>
>>> And that is the bug - you should not be initializing two languages,
>>> because you prevent the (potential) translations of either of them from
>>> being picked up, and served to the client.
>>>
>>> I can see that you considered this as "all english" - but if you think of
>>> this as separate languages, and in terms of how you read-in the language
>>> translation files, then the mistake is easy to see.
>>>
>>>
>>>> This is
>>>> way it was not specified before. This is why I am still not completely
>>>> convinced it is a good idea not to let the users be explicit.
>>>>
>>>
>>> You are not looking at this in the right way; you are wrong - look in
>>> terms of your design, and it should be immediately clear.
>>>
>>> For example, think about setting "default language" as 'it' and 'es' ---
>>> and try to walk thru the logic in gluon/languages.py - then it should be
>>> very clear that only _one_ language should be initialized.
>>>
>>> After that point, you can extend this to see that adding a non-country
>>> specific language to the initialization does not cause any bad behavior, and
>>> can be useful (help deliver the language appropriately more often).
>>>
>>> Just remove 'en-uk' from this patch, and it will be fine.
>>>
>>> - Yarko
>>>
>>>>
>>>> On Nov 24, 12:14 pm, Yarko Tymciurak <resultsinsoftw...@gmail.com>
>>>> wrote:
>>>> > On Mon, Nov 23, 2009 at 2:25 PM, mdipierro <mdipie...@cs.depaul.edu>
>>>> wrote:
>>>> >
>>>> > > Yarko's patch is tentatively in trunk since nobody seems to complain
>>>> > > about this change in behavior.
>>>> >
>>>> > You made an error with the change you made in this patch: you added
>>>> 3
>>>> > languages, 'en', 'en-us', and 'en-gr';
>>>> > This should only be either 'en-us' (the language of the distro), or at
>>>> most
>>>> > ['en-us', 'en'].
>>>> >
>>>> > As you've done it, you've introduced another bug.
>>>> >
>>>> > Putting en-gr will prevent 'en-gr' from being seen if it is a
>>>> translation
>>>> > file UNLESS application FORCES a base language (for example).
>>>> > This means that 'behavior' and 'behaviour' will not be appropriately
>>>> > picked up from a languages/en-gr.py file UNLESS EACH application
>>>> forces
>>>> > language to 'en-us' (or some other, non-[en-gr] language).
>>>> >
>>>> > For example, a 'en-us' app will NOT be able (with this app) to
>>>> correctly
>>>> > display to someone in England, who has their language set as 'en-gr'.
>>>> >
>>>> > Please fix this in trunk: to ['en-us']; ['en-us', 'en'] would also
>>>> work
>>>> > appropriately and be acceptable.
>>>> >
>>>> > - Yarko
>>>> >
>>>> > > Massimo
>>>> >
>>>> >
>>>>
>>>>
>>>
>>
>>
>>
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"web2py-users" group.
To post to this group, send email to web2py@googlegroups.com
To unsubscribe from this group, send email to
web2py+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/web2py?hl=en
-~----------~----~----~----~------~----~------~--~---
