I want to separate the discussion for the design for the
internationalization of ftview/ftstring and C99 issue, so please let me
change the Subject.
(2016/09/23 10:34), Alexei Podtelezhnikov wrote:
Is it bad idea
to add an option to take the sample string from stdin
or other file?
I don't have Japanese or Chinese text handy, but I want it on my finger tips
for Freetype testing. What is bad about hard coding it into Freetype ?
Of course, there would be many developers who want to automate the
testing and they want to have the collection of the short texts for the
testing. I agree.
But it would be another option to separate the binary executable and the
text files, aslike gettext framework on POSIX environment (which only
includes default - usually locale "C" - messages in the binary
executable, and other languages are separated as message catalog file).
I'm not suggesting to change ftview/ftstring to use gettext framework
(because it would be too much), but having the binary executable and the
collection of text files for testing would not be so difficult solution
for the developers automating the tests. In my personal impression, for
some languages, it would be difficult to define the best sample text, so
it would be difficult for us to choose the best text which is worthful
to hardwire into the binary executable.
For Japanese, the earliest typeface specimens used the quote of "iroha"
for hiragana/katakana (いろはにほてと...), and the quote of "Thousand
Character Classic" (
https://en.wikipedia.org/wiki/Thousand_Character_Classic , "天地玄黃宇宙
洪荒...") for Hanzi. However, Japanese texts are usually the mixtures of
kana & Hanzi, the post-WWII specimens used different texts mixing both
characters, like, "愛のあるユニークで豊かな書体".
Freetype-devel mailing list