On 10 Nov 1999 04:03:47 +0100, Lars Gullik Bj°nnes wrote:
>One other thing we ofcourse can to is to just use _ for all chars that
>is not in the keep list already, and drop the 0x7f completely. That
>actually has some merits...
>
>| The reason for isalnum() in 1.0.4 and 1.1.1 was precisely to avoid such
>| situation.
>
>The problem with isalnum is that is disallows a lot of chars otherwise
>perfectly ok.
When I first introduced the idea with isalnum() I considered this. And
I chose the 'fundamentalist' way to explicitly allow non-alnum chars,
to be on the safe side instead of using isgraph(), which too heavily
depends on locale support when used on non-Unix systems: So you never
know which hex char is interpreted as a glyph and which one as a
control char. Or do you know which encodings are used in arab countries
or, say, in Hungary or Bulgaria on non-Unix file systems.
>| LaTeX can handle 0x8f all right if the pool file is created that way.
>| There are quite a few emTeX users who configured TeX compiler for CP437
>| rather than for ISO8859-X (And they are happily running TeX on file
>| system of CP437 character file names.)
>
>And we now use ascii only.
How can you be sure without '& 0x7f' for all the numerous different
Double Byte encodings, commonly used for PC file systems in asian
countries? Perhaps some of the asian language porters could give a
helpful summary of the situation?
Greets,
Arnd
