On 21 Jul 1999 07:18:49 +0900, [EMAIL PROTECTED] wrote:
>This is not your code, but I don't like this one, too:
>> LString SpaceLess(LString const & file)
>[...]
// Substitute chars that LaTeX can't handle with safe ones
>> name.subst(';', ':');
Yes, according to Murphy this leads to a vicious name handling circle:
name.subst(';', ':'); <-]
[...] ^
somewhere.C: |
name.subst(':', ';'); ->]
If OnlyFilename() works correctly, I'd propose then
name.subst(';', ',');
>As for 8 bit characters, it may be more efficient to provide
>LString::discardSign() which converts all the occurence of 0xa0-0xff
>to 0x20-0x7f which must of course be called after 0xaf, 0xba, 0xbb
>and 0xdc are converted *AND* before these already existing special
>characters handling.
>(Well, on the second thought, 0x80-0x9f are used in the codepages.)
So, must we write a clumsy substitution test list with a hundred
entries or so? This '// Substitute chars that LaTeX can't handle'
problem should be present on most if not all OS!
And what about latin2-x?, not to speak about non-latin encodings... Oh
boy, oh boy...
Thanks to all those patient devvies,
Arnd
