harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stepan Mishura" <stepan.mish...@gmail.com>
Subject Re: [classlib][text]Different impl of DateFormat/SimpleDateFormat between RI and ICU
Date Mon, 28 Jan 2008 08:59:07 GMT
Hi Tony,

Thanks for the analysis. As I understood you suggested to record most
of differences as non-bugs but what about obeying the spec? For
example, I've reviewed the first case.

According to the spec. for SimpleDateFormat:
...
Letter:                                   Z
Date or Time Component: Time zone
Presentation:                       RFC 822 time zone
Examples:                            -0800
...

RFC 822 time zone: For formatting, the RFC 822 4-digit time zone
format is used:
     RFC822TimeZone:
             Sign TwoDigitHours Minutes
     TwoDigitHours:
             Digit DigitTwoDigitHours must be between 00 and 23."

For me RI's behaviour fits for the spec. and ICU's doesn't. So it is
not clear why it should be marked it as non-bug difference (we agreed
to mark a difference as non-bug if the spec is not clear or silent and
RI's behaviour is illogical.). Did I miss something?

Thanks,
Stepan.

On 1/28/08, Tony Wu <wuyuehao@gmail.com> wrote:
> Hi, all
>
> I collect the differences I met here, let's discuss them case by case.
>
> 1. For pattern "ZZZZ", ICU returns GMT-07:00 but RI returns -7000. For
> pattern "z, zzzz", ICU returns "GMT+14:00, Line Islands Time" but RI
> returns "LINT, Line Is. Time" in "Pacific/Kiritimati".
> ....
>
> And lots of things like that, also result in some failures in MessageFormat.
> I suggest to correct our testcases and record them as non-bug
> differences now and fix them one by one in the future if any product
> were blocked.
>
>
> 2. ICU supports extra pattern chars, like ;Y', 'u', 'e', etc.  refert
> to http://www.icu-project.org/apiref/icu4j/com/ibm/icu/text/SimpleDateFormat.html.
>
> I think it's a value-add for us, let's report it as a non-bug difference.
>
> 3. ICU has different explanation on the pattern 'y'. It may result out
> in different output, such as 2008 on ICU but 08 on RI. I suggest to
> keep it as is unless we have to fix.
>
> Any other difference you found, please let me know :)
>
>
> --
> Tony Wu
> China Software Development Lab, IBM
>

Mime
View raw message