> "publ.lib.ru", cache=0x1) at htslib.c:4561
> #2  hts_dns_resolve_ (opt=opt@entry=0x7fffdc023840,
> _iadr=_iadr@entry=0x7fffe419f970 "publ.lib.ru",
> addr=addr@entry=0x7fffe418e7e0,
> error=error@entry=0x7fffe418e7d0)
>     at htslib.c:4683
> #3  0x00007ffff6529d94 in hts_dns_resolve2
> (opt=0x7fffdc023840, _iadr=0x7fffe419f970
> "publ.lib.ru", addr=0x7fffe418e7e0,
> error=0x7fffe418e7d0) at htslib.c:4725
> #4  0x00007ffff652a872 in newhttp (opt=0x7fffdc023840, 
> _iadr=0x70 <error:
> Cannot access memory at address 0x70>, 
> retour=0x7fffe251dc50, port=80,
> waitconnect=0) at htslib.c:2150

Humm, the backtrace seems unfortunately unreliable: _iadr is valid in the
callee, not in the caller (with a similar 0x70 address), but this crash has
already been seen on Android, without much success yet unfortunately.

Did the crash occur after a while, or immediately after starting the mirror ?