Jump to content

llen_llen

Community Members
  • Posts

    7
  • Joined

  • Last visited

llen_llen's Achievements

Tiro

Tiro (1/14)

1

Reputation

  1. Thank you! It worked for me perfectly! It seems that -mod=modwas the critical part, after all. Sorry, but I'm not sure about what you meant in this part.
  2. I followed instructions on the following page: http://trac.wildfiregames.com/wiki/Adding_font_support except for some typos and directory delimiter difference seen in the example. I think Kanetaka too. EDIT: As for me, I had problem installing the game in the first place. Extracting of public.zip kept aborting at 64% whenever I tried, so I decided to decompress the exe and put every file directly into the target folder.
  3. I had my post blown away for some reason so I had to rewrite this from scratch Isn't there any way to save my draft...? Care to elaborate? I really ought to have First, by "the current style" I mean the way we pre-render all glyphs we use into one sprite image and load it for displaying. This means an extra letter costs more memory than that of in a ttf file, particularly for East Asian characters which tend to be bulky than alphabets. To dump whole character set from an all-in-one font such as Hanazono or MicroHei results in enormous textures. Meanwhile, many standardized language-specific subset(s) are known to suffice daily usage, which typically contains some thousands of glyphs (it sounds still big, but remember current official font packs contain some tens of thousands of them). We can choose ones for each locale mod so that substantially reduce textures' size. This is the general idea. Yes, you are right for the most part, and I truly appreciate your effort to make them look more native. The only problem here is too many characters. The most of them are actually dispensable locale-wise, unless you take really seriously of interoperability, such as, you must be able to chat every character across locales or something. Sure! The first choice for all Asian locales will be Source Han Sans family today, which contains locale-specific variations and 7 weights, under Apache 2.0 License. Our community build (see below) uses it too. Other qualified fonts include: Japanese - M+ (sans-serif; 7 weights) with many derivatives (Mgen+, Rounded M+, LogoTypeGothic...), Sawarabi (sans/serif), Nukamiso (sans), GT(sans/serif/script) etc. Chinese (Taiwan) - official CNS 11643 fonts (script/serif; Trad.), Arphic 1 2 (script/serif; Simp./Trad.), cwTex family (Trad.) etc. Chinese (China) - only WenQuanYi family (includes MicroHei) found free with explicit license. Note that simplified Chinese font created in Taiwan may be inaccurate in detail, and vice versa. Thank you. Kanetaka and I are trying to make a lightweight community build, here are some screenshots: Thanks to the official fontbuilder script, we were able to optimize font size and hinting, so it looks much better. But for some reason we are currently unable to build a correctly cached mod. Could you examine this original file? https://www.dropbox.com/s/gqcvwia93s9w46y/jplocale%28uncached%29.zip?dl=0 (~30MB; zip) And I believe that simplified/traditional Chinese font packs could be built in similar way. That's okay. I was just curious if font format will change in the near future.
  4. Thank you for providing font support! However I believe East Asian font packs need proper subsetting as long as we carry on the current style. There are several reasons: The texture size is directly reflected in memory usage, so you'll suffer from severe memory overuse with huge textures. See http://trac.wildfiregames.com/ticket/2843 A great portion of characters actually only exist for (historical|regional) (terms|proper names) otherwise of no use (unless for translating Asian faction mod etc). A "needed" character set for China/Taiwan/Japan is different from other ones by thousands of characters, due to linguistic/orthographical difference (not to mention Korea...). Covering all of them only adds a large number of unusable letters to every language.At least Japanese community can offer a more dedicated JP font pack. How do you like it, leper? P.S. Aside from chinese characters, I don't think the current implementation can handle Arabic or Indic alphabets while Persian translation is ongoing. Is there any plan to replace the architecture?
  5. Thank you, GunChleoc and leper. That helped me a lot to understand what's going on. I would prefer to group them into with- and without-variable forms. (singular) : (plural)#1: "%(arrowcount)s arrow / second", "%(arrowcount)s arrows / second"#2: "%(arrowcount)s arrow / %(time) seconds", "%(arrowcount)s arrows / %(time) seconds"And in my opinion, the slash and "second(s)" are better not to be separated, as well as prepositions and following nouns.
  6. A problem with a recently added string (which I'm not sure here is the right place to ask about): string #2890 on Transifex "second"/"%(time)s seconds" contains a pair not only different in pluralization, but in an additional variable. This makes translation unable for non-plural languages, such as Japanese, because Transifex doesn't allow them to separate. Is it solvable by changing the original string table or something?
  7. Has anybody mentioned Athenian Marine's name before? I think it should be "Epibates Athenaikos" instead of "Epibastes Athenaikos". Source: Perseus
×
×
  • Create New...