Craig A. Berry <[EMAIL PROTECTED]> writes:
>
>
>Many thanks.  Actually, when I run this I get the correct output 
>(acf1adf1aef1aff1b0f1b1f1b2f1b3f1) and no warnings, which I think confirms 
>that we have a buffer getting corrupted (or lost our reference point in it? 
>[or separate I/O operations getting the two halves of one character?]) 

I think the Encode issue is due to the last case - a ksc5601 character 
spanning two stdio buffers - 1st octet in one, 2nd octet in the other.
I am hoping that problem is that ungetc() is loosing an octet 
(why is that not noticed in other tests) or perhaps that multiple-ungetc
re-orders the octets.

>rather than trouble parsing some particular legitimate byte sequence.

Agreed.

>
>If I create a file containing only lines 229-231 from ksc5601.enc, I also do 
>not get the warning when reading from it that I get when reading the entire 
>file.

Again good - Encode is blameless IO is looking guilty ;-)


-- 
Nick Ing-Simmons
http://www.ni-s.u-net.com/



Reply via email to