topic Re: _io_tfs_install and const char * in MQX Software Solutions

_io_tfs_install and const char *

razed11 — Wed, 10 Sep 2014 18:26:16 GMT

Does this routine modify the first argument? If not why is it not const?

_io_tfs_install(TFS_ROOT, tfs_data);

Thanks,

Kenny

Re: _io_tfs_install and const char *

danielchen — Fri, 12 Sep 2014 02:34:37 GMT

Hi Kenny

Please see the function _io_dev_install_ext(char *identifier,....),

you can see this line "if (identifier[i] == IO_DEV_DELIMITER)". if identifier is const ， maybe there will be a compiler error.

/* Search for delimiter */

for (i = 0; i < IO_MAXIMUM_NAME_LENGTH; i++) {

if (identifier[i] == IO_DEV_DELIMITER) {

found++;

} else if (identifier[i] == '\0') {

break;

} /* Endif */

} /* Endfor */

Have a great day,
Daniel

Re: _io_tfs_install and const char *

razed11 — Fri, 12 Sep 2014 19:05:31 GMT

Thanks Daniel. This doesn't really answer the question. I understand that it will create a compiler warning. I'm suggesting that if the routine doesn't modify the value that the interface should declare identifier as const. That allows me to use:

static const char *IDENTIFIER = "tfs:";

Otherwise I have to copy this to a buffer and it is clumsy for the client. A string literal will not create a compiler error even though it is technically const or at least not ensured to be. That is for historical reasons.

If the interface is changed to const char * then it is backward compatible. Using a literal is not good practice in my opinion especially if it must be used more than once.

I suppose this is less of a question and more of a feature request.

Kenny

Re: _io_tfs_install and const char *

danielchen — Fri, 12 Sep 2014 19:57:48 GMT

I understand, I will check your request with developers