can't convert mixed encoded file

[ifq]

I have lots of C source file, which mixed with GB2312 and UTF-8 comments. I want to convert all GB2312 to utf8, and recode

$ recode -v gb2312..utf8 io.c
Request: EUC-CN..:libiconv:..UTF-8
Shrunk to: EUC-CN..UTF-8
Recoding efmp_io.c... failed: Invalid input in step `EUC-CN..UTF-8'

If I manually delete utf8 encoded comment in io.c file, then the process goes well.

Is it possible to add an option, like -i utf-8 , which keep specific encoded part in file, and convert other part of the file.

I searched a lot today, seems like no tools support this kind of function.

[rrthomas]

This is in general not possible (how would recode know which bits are in which encoding?).