Hi, Jan and Peter:
Thanks for your reply.
I am not sure if my understanding about SUNW_LOC is correct:
- Add/Remove of locales are done with a granularity of region.
- SUNW_LOC of each package will define what region the package belong to
- SUNW_LOC may contain several different locales, say LA, LB, LC for one
region R at same time
- As long as (at least) one locale of LA/LB/LC is specified, the package
will be installed for region R.(If RE forgot to put LC there, it doesn't
matter, does it?)
- There won't be a package shared by 2 or more regions.
If there are no such packages, then the following paragraph can be dropped.
- Considering a package shared by 2 regions, its SUNW_LOC should contain
LA1/LB1/LC1 and LA2/LB2/LC2 for region R1 and R2 respectively.
- Let's presume the package is new one from community -- we don't have a
previous reference. If RE forgot to set SUNW_LOC to LA2(or LB2 or LC2),
then the package can be installed for region R2 as well.
- If RE forgot to put all of LA2/LB2/LC2, then the package won't be
installed for region R2.
For the last case, we won't be able to detect the error until PCT or SCT
> Peter.Nugent@xxxxxxx wrote:
>> Harry Fu wrote:
>>> 1) Is there a way that a user could know --
>>> * what packages are installed on the system for a specific locale?
>>> * what locales are installed on the system for a specific package?
>>> I am asking it because it seems that the design relies on SUNW_LOC
>>> field provided by package owners very much. However, I think it is
>>> likely that some human mistakes may introduce bugs that some packages
>>> are missed. Then when a user chooses to add a region, those packages
>>> are not installed. In that case, is there a way to identify the issue
>>> without extensive testing?
>> localeadm relies on SUNW_LOC being correct and up to date and I feel
>> this is the correc tthing to do. If it is not then that is a bug in
>> the relevant pkginfo file that needs to be fixed.
>> Damien did a comparison recently of packages per geo/locale in the
>> current localeadm config file and locales listed in SUNW_LOC and AFAIK
>> they were pretty close and he filed 1 or 2 bugs against the relevant
>> pkgs where there were missing entries in SUNW_LOC. I don't recall what
>> the status is but Damien can tell us more. Thanks for reminding us
>> about this issue.
> Just a comment about verification of SUNW_LOC data. The onepager
> mentions an automated test which will compare new and old localeadm
> package lists.
> So, an extensive manual testing will not be necessary here.
i18n-discuss mailing list