This is mostly a cosmetic thing for consistency with the other
languages, but it also works around an unfortunate line wrapping
problem: the line "2.Une fois que vous disposez d'un KI version 2.0"
gets wrapped at the version number, between "." and "0". This is
probably because it interprets the "." as an end of sentence, and there
seems to be no good way to prevent this (we can't use a non-breaking
space, because we don't want a space). By adding the spaces in the
numbered lists, the problematic line becomes long enough that the entire
version number gets wrapped onto the next line, as expected.
In French, some punctuation characters (question mark, exclamation mark,
colon, semicolon) also have a space before the symbol, not just after
the symbol as in most other languages. This preceding space should be a
non-breaking space, so that the punctuation is not wrapped onto a new
line separate from the preceding word.
The French translation was inconsistent about its quote style. Some
parts used guillemets, but others used double quotes. As far as I know,
guillemets are the standard quote form in French.
Note that in French, there are spaces on the inside of the guillemets as
well, not just on the outside as in most other languages. These inside
spaces should be non-breaking spaces, so that the guillemets are
line-wrapped together with the text being quoted.
This round-trips the entire Cyan localization database through the
plLocalizationEditor built from H-uru/Plasma#958. This means that:
- the XML files are re-encoded from UTF-16 to UTF-8 and therefore should
load faster due to less run-time re-encoding
- journals should be written out as CDATA with the esHTML data visible
in the clear
dgelessus
Adam Johnson