But your patch changes the line to:
```
#if (OPENSSL_VERSION_NUMBER >= 0x10001000L) && !defined(LIBRESSL_VERSION_NUMBER)
```
So it always falls to the error case.
> Its definitively too easy to make such mistakes which could be really bad.
This is one such mistake, unfortunately. :(
Can this ticket be assigned to whoever does the Qt stuff for y'all?
Ok, the TLSv1_{1,2}_{client,server}_method are available in libressl and thus a good number of patches are wrong. I [stripped them from the patch](http://misc.voidlinux.de/libressl-compat.patch) in a first round. 41 patches left.
Patches validation
- [x] TLSv1_{1,2}_{client,server}_method available in libressl; related patches removed
- [x] EC_curve_nist2nid not available in libressl; patches must remain
- [ ] SSL_{set,get}_ex_data, SSL_get_ex_new_index are available in libressl.
Need to closely inspect the related patches.
- [x] SSL_set_psk_{client,server}_callback not available, i.e. related patches must remain
or I could perhaps use -DOPENSSL_NO_PSK
- [ ] more to come...