Date: Wed, 19 Dec 2007 12:00:11 GMT From: Wei-Hao Syu <whsyu@ntu.edu.tw> To: gnome@FreeBSD.org Subject: Re: ports/118481: big5-2003 in converters/libiconv Message-ID: <200712191200.lBJC0BBY076882@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR ports/118481; it has been noted by GNATS. From: Wei-Hao Syu <whsyu@ntu.edu.tw> To: Alexander Nedotsukov <bland@FreeBSD.org> Cc: bug-followup@FreeBSD.org Subject: Re: ports/118481: big5-2003 in converters/libiconv Date: Wed, 19 Dec 2007 19:57:48 +0800 Why we want to replace big5 instead of directly use big5-2003? Because =20= in some applications you can only choose big5, no options for =20 big5-2003, you have to hack each of them if you want. Some =20 applications read locale info to convert texts to original big5, and =20 again it's not easy to modify each of these program to use big5-2003. Please note that this hack is an option, and I don't think it should =20 be enable by default. If someone know what he need and don't care about the incompatibility =20= part of them, then just enable it. Otherwise everything works just as =20= before. For the compatibility issue, I think that you are talking about the =20 0xA140-0xA2CE mapping. big5-2003, big5-1984, and cp950 have different mapping in some symbols =20= at this section. (but with almost same look, such as 0xA156 -> U+2013/U+2015) This =20 issue is already exist because some big5 data generated by windows =20 system, we assume it's big5-1984 but actually it is cp950. I agree with you that we should also ask GNU libiconv team to replace =20= the original one which comes from obsolete unicode 1.1 table. (another =20= issue of this one: big5-1984 does not define katakana/hiragana, but =20 this table has an incorrect one in it.) On 2007/12/17, at =A4U=A4=C8 1:37, Alexander Nedotsukov wrote: > I understand the difference. My question was more about why you can =20= > not use BIG5-2003 when it appropriate? The point is BIG5 in its =20 > original form is not a subset of BIG5-2003. There are code points =20 > defined differently. So what you asking for is technically illegal. =20= > However if this is an *official* way you do it in Taiwan (which will =20= > be really weird case) please convince GNU libiconv developers to =20 > switchover. > > Wei-Hao Syu wrote: >> because of katakana and hiragana. >> >> The major difference between big5-2003/big5-1984 is big5-2003 has =20 >> katakana and hiragana mapping ( from big5-eten). big5-2003 is part =20= >> of official standard in Taiwan and most Taiwanese have the =20 >> requirement (katakana, hiragana support) when using ftp/bbs with =20 >> big5 encoding, that is why we need this one. >> >> On Thu, 13 Dec 2007 11:25:35 +0900 , Alexander Nedotsukov = <bland@FreeBSD.org=20 >> > wrote: >>> Could you explain why you need this hack and what is more =20 >>> important how it will interact with other variants of BIG5 family, =20= >>> please? I can see that this silent switchover may lead to =20 >>> incompatibility between hacked and clean systems which is not good =20= >>> thing IMHO. In any case I strictly recommend you to put pressure =20 >>> on GNU libiconv developers to resolve issue at the right place. >>> >> >> >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200712191200.lBJC0BBY076882>