[Bug time/23758] Improve the width of alternate representation for year in strftime

[tamuki at linet dot gr.jp]

https://sourceware.org/bugzilla/show_bug.cgi?id=23758

--- Comment #2 from TAMUKI Shoichi <tamuki at linet dot gr.jp> ---
The previous patch adds a way to override the padding, when %EY is
exanded using locale data, rewriting the %Ey contained in the format
string.  That needs for new memory allocations.  So, I revised the
patch to be handled by not rewriting the format string.

Along with this, the modified description is shown below:

----------------------------------------------------------------------
The Japanese era name is scheduled to be changed on May 1, 2019.
Prior to this, improve the width of alternate representation for year
in strftime to prevent the trouble to be mentioned below we saw in the
past from becoming obvious again from the year after the era name
changes onward.

In one-digit of the Japanese era year such as Heisei year 2 (1990),
the conversion specifiers %Ex, %Ec to be included in the format
specification of strftime, are currently converted as follows.
However, since the width of alternate representation for year of the
output string is different from the others, there was a sense of
incompatibility.

    "%Ex"       : "平成2年08月04日"
    "%Ec"       : "平成2年08月04日 08時06分02秒"

These conversion specifiers are implemented by combining multiple
conversion specifiers.  In ja_JP locale for example, the combinations
of conversion specifiers are defined as follows.

    "%Ex"       : "%EY%m月%d日" (era_d_fmt)
    "%Ec"       : "%EY%m月%d日 %H時%M分%S秒" (era_d_t_fmt)

    "%EY"       : "%EC元年" | "%EC%Ey年" (era_format)
    "%EC"       : "平成" | "昭和" | "大正" | "明治" (era_name)

Currently, the width of the output string of the numeric value of
conversion specifier %Ey is not padded.  By changing the width padding
with zero to 2, the conversion specifier %EY is converted as follows,
and then the expected results are also obtained for the conversion
specifiers %Ex, %Ec.

    "%EY"       : "平成2年" -> "平成02年"

    "%Ex"       : "平成02年08月04日"
    "%Ec"       : "平成02年08月04日 08時06分02秒"

Since only one Japanese era name is used by each emperor's reign, it
is rare that the year ends in one digit or lasts more than three
digits.  In addition, the width of month, day, hour, minute, and
second is 2, so adjust the width of year the same as them, and then
the whole display balance is improved.  Therefore, it would be
reasonable to change the width padding with zero of %Ey default to 2.

Currently in glibc, besides ja_JP locale, the locales using the
conversion specifier above are lo_LA (Laos) and th_TH (Thailand).  In
these locales, they use the Buddhist era.  The Buddhist era is a value
obtained by adding 543 to the Christian era, so it is not affected by
this change of the conversion specifier %Ey.

Furthermore, for the output string of the conversion specifier %EY,
an optional flag is given to the conversion specifier so that it can
be also used the current non-padding format, and the padding format
can be controlled.  To achieve this, when an optional flag is given to
the conversion specifier %EY, the %Ey included in the combined
conversion specifier is interpreted as if decorated with the
appropriate flag.

    "%EY"       : "平成02年"
    "%_EY"      : "平成 2年"
    "%-EY"      : "平成2年"

As an application example, by combining conversion specifiers, it can
be converted into the following display format.

    "%Ex(%a) %X"                        : "平成02年08月04日(土) 08時06分02秒"
    "%_EY%b%e日(%a) %_H時%_M分%_S秒"        : "平成 2年 8月 4日(土)  8時 6分 2秒"
    "%-EY%B%-d日(%a) %-H時%-M分%-S秒"       : "平成2年8月4日(土) 8時6分2秒"

-- 
You are receiving this mail because:
You are on the CC list for the bug.