On 09/06/2010 09:00 PM, Stefan Fritsch wrote: > On Monday 06 September 2010, Ruediger Pluem wrote: >>> + >>> + /* >>> + * The apr-util docs wrongly states encoded strings are not >>> 0-terminated. >>> + * Let's be save and allocate an additional byte. >>> + */ >>> + len = 1 + apr_base64_encode_len(sizeof(id)); >>> + encoded = apr_palloc(r ? r->pool : c->pool, len); >>> + apr_base64_encode(encoded, (char *)&id, sizeof(id)); >>> + encoded[11] = '\0'; /* omit last char which is always '=' */ >> Why 11 and not len? > > I didn't want to depend on wether apr_base64_encode_len() includes > the terminating \0 or not. But as I am APR committer now, too, I > should probably fix the docs instead. > > Besides, it would have been encoded[len-3] or encoded[len-2] which > is IMHO even less readable than encoded[11].
Given your latest comment (that you do not need the '=') IMHO encoded[len-3] or encoded[len-2] is more readable then encoded[11]. > > I have commited all your other suggestions. Thanks. Regards Rüdiger
