> From: owner-openssl-us...@openssl.org On Behalf Of ml
> Sent: Saturday, 26 May, 2012 16:18
> I try to realize as base64 encoding
> char *base64(char *input, int length)
> {
> BIO *bmem, *b64;
> BUF_MEM *bptr;
>
> b64 = BIO_new(BIO_f_base64());
> bmem = BIO_new(BIO_s_mem());
> b64 = BIO_push(b64, bmem);
> BIO_write(b64, input, length);
> BIO_flush(b64);
> BIO_get_mem_ptr(b64, &bptr);
>
> char *buff = (char *)malloc(bptr->length+1);
> memcpy(buff, bptr->data, bptr->length);
> buff[bptr->length] = 0;
>
> BIO_free_all(b64);
>
> return buff;
> }
> but it does not work can you help me
>
It works for me, with the needed #include's and call added.
What exactly is your problem? What did it do, what did you
expect, and how do those differ?
I note the cast of malloc to (char*) is suspicious.
If you #include'd <stdlib.h> as is required to correctly
declare malloc, no cast is needed in C since about 1989.
A cast is needed in C++, ut in C++ you should normally use
(typesafe) new and delete rather than malloc/free/etc.
And a call to malloc *without* #include'ing <stdlib.h>
is not allowed in either current C (C99 or later) or C++,
but the use of a declaration after executable statements
within a block, as you have, is *only* allowed in those.
Also you should always check if malloc/calloc/realloc
returned a null pointer (failed) before using it. In some
applications you can actually handle the error usefully,
and even when you can't a clear and specific assert() or
similar suicide is better than an unexplained fault.
(Also fopen, tmpnam, getenv, etc, etc.)
If you are using some sort of nonstandard C (or C++, or even
some other C-like something), give details; it may or may not
be sufficiently compatible with C as used for OpenSSL.
______________________________________________________________________
OpenSSL Project http://www.openssl.org
User Support Mailing List openssl-users@openssl.org
Automated List Manager majord...@openssl.org
