ok stsp@

gotd and gotwebd are using almost the same log.c file, so unify. This only slightly changes gotwebd logs to include the process name, like gotd. ok stsp@

Now that `listen' is a top-level, `unix_socket_name' is useless. The default socket remains /var/www/run/gotweb.sock, to listen on an alternate socket now `listen on socket "/some/other/sock" can be used. part of a larger diff that's ok stsp@

gotwebd only uses the server name (i.e. the Host HTTP header) to match the server blocks. Since gotwebd by design sits behind an http server which is expected to filter virtual hosts, there's little point in having `listen' per-server. part of a larger diff that's ok stsp@

- don't assume an implicit "listen on socket" when another listen was specified in a server block - allow servers to share the sockets they're listening on - fix unix_socket_name; now it's actually used - remove `listen on socket off' and `unix_socket off' options that are now no longer useful - unify the internal handling of the sockets, i.e. don't consider unix socket special. ok tracey@

h is free'd by add_addr in case of error, but we still leak the getaddrinfo result. while here make the check strictier: addr_addr() fails with -1.

Unlike max_repos_display which limits the number of items per page, max_repos is an hard-limit on the number of repos gotwebd would process. "kill it with fire" tracey@ about the idea ok stsp

ok jamsek

proc.c really shines when there's a network of different types of processes, potentially with a various number of instances each, that needs to exchange messages. Gotwebd instead has a much simpler design, and using proc.c causes more overhead (/headaches) than it solves. So, this attempts to provide the same functionalities but with a much simpler implementation that fits gotwebd better. ok stsp@

Instead of hardcoding the flags, keep ai_{family,socktype,protocol} from the getaddrinfo() call and apply them here. ok stsp@

It was added due to gotwebd weird structure sockaddr_storage handling. Instead, save the size reported by getaddrinfo() and not reach into the struct sockaddr_storage at all (except for extracting the port number for diagnostics purposes.) sockets_conf_new_socket_fcgi() gets an hardcoded ipproto to zero (which is the only value it can get in practice, and keeps for the moment the hardcoded SOCK_STREAM. It'll be cleaned in a follow-up. ok stsp@

This changes how we handle the port number: bubbles up the local portstr added in previous commit and lets getaddrinfo() deals with port numbers and services name. getservice() can be gc. While here add the missing free() in parse.y. ok stsp@

ok plus tweaks stsp@

ok stsp@

ipproto is always -1, so delete the dead code. ok stsp@

in 1a0c81fb61 I've added a range check too strict for max_repos_display, as zero is allowed and means "show all the repositories".

saves a bit fiddling around with ternary operators. sockets_run() will reset the chroot to the default since it calls again config_init(), but at that stage it's not important anymore. ok stsp@

Supporting both "no" and "on" as valid syntax with an opposite meaning could easily result in accidental misconfiguration. Suggested by stsp, tracey agrees. Patch by Josiah Frentsos, thanks!

ok stsp

While they're the same value, it's more correct to use AF_UNSPEC in this case. While here, change SOCK_DGRAM to SOCK_STREAM since it's what we'll be using. ok stsp@

host_dns() (now just host()) uses getaddrinfo() which knows already how to handle IPv4/IPv6 addresses. Saves a good amount of duplication. ok stsp@