Re: [PATCH 08/11] Add a buffer size parameter to convert_utf8_to_gsm

[Denis Kenzior <denkenz@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 959 bytes --]

Hi Andrew,

> This is needed when encoding a string into gsm that needs to fit a
> in a given number of bytes.  Just taking the first N bytes of the
> resulting string may leave you with byte 1 of a two-byte character at
> the end of the buffer.  As far as I can tell there's no easier way to
> avoid that.

In gsm it is actually pretty easy, the escape character for two-byte sequence 
is always 0x1b, so chopping a string up could be done quite easily.

>  unsigned char *convert_utf8_to_gsm_with_lang(const char *text, long len,
>  long *items_read, long *items_written, unsigned char terminator,
> +				long max_size,
>  				enum gsm_dialect locking_shift_lang,
>  				enum gsm_dialect single_shift_lang);

If there's still need for this, I'd prefer we reuse long *items_written as an 
in/out argument for the max size or add another function entirely, adding yet 
another int argument is getting a bit ridiculous.

Regards,
-Denis