On Jan 24, 2018, at 1:47 PM, Warren Young <[email protected]> wrote: > > On Sep 16, 2017, at 1:27 PM, Andy Goth <[email protected]> wrote: >> >> Indeed, 1.1.0f is a version which includes the BIO_ADDR type, which is a >> union containing struct sockaddr_in6 among others. I don't think there's any >> question Fossil is trying to read an IPv6 address structure as if it were >> IPv4. > > Has anyone tried addressing this? I still see it on my Raspberry Pi running > a recent version of Raspbian.
In case it’s not clear, I mean that I see it with the tip of trunk. [1cbaf3bdd9] lsb_release -s says, "Raspbian GNU/Linux 9.3 (stretch)” > Since I haven’t seen it on any of my 64-bit hosts, I’m guessing it has > something to do with a peculiarity on 32-bit Linux. I’ve poked at it some, and it’s only happening in the HTTPS code path. If I clone http://fossil-scm.org on the Pi, it shows the correct IP, 45.33.6.223. I tried to get GDB to dump the contents of *iBio, but it complains that it’s an incomplete type. I’m guessing I need debug libraries for OpenSSL, but Googled instructions for installing them don’t work. (“apt install libssl-dbg.”) About the only constructive thing I can add is that it looks like the call on like 393 should probably be a call to BIO_set_conn_ip() rather than to the underlying function. That’s mainly cosmetic, but I found that Googling on the underlying function call was less helpful, since it’s intended to be an internal interface. _______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
