On 04/06/2015 12:57 PM, Hubert Tarasiuk wrote:
The problem was, that `remote_to_utf8` would allocate a new buffer, but the buffer was never freed. (It was the `new` pointer, later copied to `host`, used for `idna_to_ascii_8z`. After returning from idn_encode, it was out of scope.)
I'm not pretending to cast down your patches, I'm sure you did a thorough research, probably using Valgrind. But just being pedantic, the reason of the leak is not because *new becomes out of scope, since idn_encode() is only called at url_parse(), and the *new pointer returned by the former is catched by the latter and stored in u->host, which is in the end freed at url_free() (kinda looks like, at a glance :D).
And since the first part (`tests`) fails, `make check` does not even make it to the second part (`testenv`). (Which is probably another bug, not a feature :D.)
Probably ;-D Regards, - AJ
