Hi Roland,

thank you very much for your time and coming
with the proposal.

Since the original fix already went through
the code review process and was already integrated,
I think you would need to file new bug, if you
would like to integrate your changes into the gate.

I might be wrong in this point, Dave Miner might provide
you with more accurate information, where to start and take
a look if you would like to proceed with integrating your
changes.

Thank you,
Jan



Roland Mainz wrote:
> jan damborsky wrote:
>> William James wrote:
>>> On Tue, Apr 8, 2008 at 4:30 AM, Dave Miner <Dave.Miner at sun.com> wrote:
>>>> Jan Damborsky wrote:
>>>>  > Hi Sarah,
>>>>  >
>>>>  > I have done changes according to the Dave's and Frank's
>>>>  > suggestions:
>>>>  >
>>>>  > * timeout not implemented
>>>>  > * implemented UI complies with the mock up
>>>>  >
>>>>  > Could you please look at the updated webrev ?
>>>>  >
>>>>  > It is available at
>>>>  >
>>>>  > http://cr.opensolaris.org/~dambi/bug-973-2/
>>>>
>>>>  set_lang.bash, 41: should be "To select the desktop..."
>>> Why do you use bash instead of ksh93?
>> There was no technical reason not to use ksh93 - I assume ;-)
>>
>> As Sanjay already pointed out, this script
>> is only temporary solution and hack from
>> design point of view (please see bug report
>> for details) - different approach will have
>> to be worked out for next release.
>>
>> I chose the bash due to the time constraints -
>> at that point I was more familiar with bash than
>> ksh93.
>
> To end this debate - attached is a ksh93 version of the script. I did
> some minor changes in functionality, e.g. ...
> - The script now explicitly sets PATH to avoid that any weired stuff
> gets picked-up by accident 
> - The script now aborts on any internal error, e.g. $ set -o errexit #)
> - The code now enables the shell's "gmacs" editor mode to allow users to
> edit their inputs using the cursor keys
> - The prompt of the "read" is now localised (assuming someone writes a
> catalog)
> - The "lang_choice" is now an integer and uses the "read" builtin's
> facilities for default values
> - General cleanup per
> http://www.opensolaris.org/os/project/shell/shellstyle/
> - The script now uses the arithmetric operators (( )) , >, < etc. and
> comparisations for numberic values instead of [[ ]]/-lt/-gt
> - The code now uses printf when expanding variable values to avoid
> interpretation of any special shell characters
> - The language name/lang pairs are now stored in a compound variable
> array. Technically this should be...
> -- snip --
> typeset -a languages=(
>     ( name="1. Chinese - Simplified"    lang="zh_CN.UTF-8" )
>     ( name="2. Chinese - Traditional"   lang="zh_TW.UTF-8" )
>     ( name="3. English"                 lang="en_US.UTF-8" )
>     ( name="4. French"                  lang="fr_FR.UTF-8" )
>     ( name="5. German"                  lang="de_DE.UTF-8" )
>     ( name="6. Italian"                 lang="it_IT.UTF-8" )
>     ( name="7. Japanese"                lang="ja_JP.UTF-8" )
>     ( name="8. Korean"                  lang="ko_KR.UTF-8" )
>     ( name="9. Portuguese"              lang="pt_BR.UTF-8" )
>     ( name="10. Russian"                lang="ru_RU.UTF-8" )
>     ( name="11. Spanish"                lang="es_ES.UTF-8" )
>     ( name="12. Swedish"                lang="sv_SE.UTF-8" )
> )
> -- snip --
> ... but the ksh93 version in the Indiana demo has a small bug, therefore
> I fell-back to an alternative (and IMO much uglier) syntax (the
> ksh93-integration update1 putback will fix that). 
> - I removed the hardcoded number "12" from the script - it will now look
> at the number of entries in the "languages" array
>
> The script is attached as "set_lang.ksh" ...
>
> ----
>
> Bye,
> Roland
>


Reply via email to