Hi Judith,
excellent point! I have similar issues when formatting translated
text. I have gotten around by adding a monkey patch myself, which I
derived from gettext (another l18n plugin for rails) it works like this:
"%{firstname}, %{familyname}" % {:firstname => "Masao", :familyname
=> "Mutoh"}
-> Masao Mutoh""
Hope this helps.
Greetings,
Juergen
# call-seq:
# %(arg)
# %(hash)
#
# Format - Uses str as a format specification, and returns the
result of applying it to arg.
# If the format specification contains more than one substitution,
then arg must be
# an Array containing the values to be substituted. See
Kernel::sprintf for details of the
# format string. This is the default behavior of the String class.
# * arg: an Array or other class except Hash.
# * Returns: formatted String
#
# (e.g.) "%s, %s" % ["Masao", "Mutoh"]
#
# Also you can use a Hash as the "named argument". This is
recommanded way for Ruby-GetText
# because the translators can understand the meanings of the
msgids easily.
# * hash: {:key1 => value1, :key2 => value2, ... }
# * Returns: formatted String
#
# (e.g.) "%{firstname}, %{familyname}" % {:firstname =>
"Masao", :familyname => "Mutoh"}
#
# Note: This code was derived from gettext string.rb package and
adopted for Probono.
# We allow {} and ${} formatting, though, will be deprecated
in the future
def %(args)
if args.kind_of?(Hash)
ret = dup
args.each {|key, value|
ret.gsub!(/\%\{#{key}\}/, value.to_s)
# support deprecated formats as well, such as "${...}" and
"{...}"
if ret=~/\$\{#{key}\}/
ret.gsub!(/\$\{#{key}\}/, value.to_s)
$stderr.puts "PROBONO DEPRECATED: string format using '${#
{key}}' is deprecated, use '%{#{key}}' in the future"
end
if ret=~/\{#{key}\}/
ret.gsub!(/\{#{key}\}/, value.to_s)
$stderr.puts "PROBONO DEPRECATED: string format using '{#
{key}}' is deprecated, use '%{#{key}}' in the future"
end
}
ret
else
ret = gsub(/%\{/, '%%{')
begin
ret._old_format_m(args)
rescue ArgumentError => e
if $DEBUG
$stderr.puts " The string:#{ret}"
$stderr.puts " args:#{args.inspect}"
puts e.backtrace
else
raise ArgumentError, e.message
end
end
end
end
end
On Apr 22, 2008, at 8:06 AM, Judith Meyer wrote:
> Hello,
>
> another quick question from me: I suppose the monkey patch leads to
> interesting results when the order of %s is exchanged in a
> translation? The whole point of having %s was to allow translators to
> move these variables to whatever place they needed to be in the
> translation, so that languages like Japanese could put the verb last
> and so on. Using %s several times without disambiguation however
> assumes that the variables will always be required in the same order.
> Can this be fixed? I developed the grammar system for Cantr.net and
> there the translation system uses variables like #NAME#, #WEAPON#,
> #LOCATION# and so on.
>
> Best wishes,
>
> Judith
