diff: relayd generate an output rule for "route to"

2023-09-12 Thread YASUOKA Masahiko
Hi, After 6.9 packets passed by "route-to" started to be evaluated when output. As the result, states are created for output direction, because it is not considered about "direct server return", has some problems (eg. the state is deleted because the state tracking is failed.) relayd(8) creates

Re: ksh(1): implement p_tv() with p_ts()

2023-09-12 Thread Todd C . Miller
On Tue, 12 Sep 2023 07:49:27 +0200, Theo Buehler wrote: > While this looks like an improvement to me, this uses a new non-portable > construct in ksh. I don't know how much we care. I don't think we care. If someone wants to ports our ksh it is easy enough to supply TIMEVAL_TO_TIMESPEC if

Re: ksh(1): implement p_tv() with p_ts()

2023-09-12 Thread Todd C . Miller
On Mon, 11 Sep 2023 22:10:49 -0500, Scott Cheloha wrote: > p_tv() is identical to p_ts() in every way except for the subsecond > conversion constants. > > Better to write p_ts() once: in p_tv(), convert from timeval to > timespec and call p_ts(). OK millert@ - todd

Re: path: speed-up pkg-config

2023-09-12 Thread Marc Espie
On Mon, Sep 11, 2023 at 09:55:53AM +0200, Marc Espie wrote: > Not to pkgconf levels, but still way faster than what we had > > Updated patch from what I've shown to people, turns out the second grep > wasn't quite working. > > This does cache the set_variables_from_env shennanigans, speeding up