Paul Mackerras

> You use `fdopen()` at https://github.com/ppp-project/ppp/blob/master/pppd/plugins/winbind.c#L320 and https://github.com/ppp-project/ppp/blob/master/pppd/plugins/winbind.c#L322, but later you use `close()` at https://github.com/ppp-project/ppp/blob/master/pppd/plugins/winbind.c#L435 and https://github.com/ppp-project/ppp/blob/master/pppd/plugins/winbind.c#442. I think you should use `fclose()` otherwise you will probably leak few hundred...

I changed 3 instances of close() to fclose().

First, this needs a sign-off, see Submitting-patches.md. Secondly, it needs an explanation that is more than simply saying what the textual change is. Don't make the reader have to reverse-engineer...

@enaess is does seem like ba7f7e0 introduced a problem with the link_stats_print variable, in that now print_link_stats will only ever produce output at most once. What was the motivation for...

Not sure why the DCO check is failing...

@enaess I think the basic problem is that link_stats_print is only ever set to 1 on initialization, hence print_link_stats() will only ever produce output once (at most). However, there are...

> I don't recall correctly, but didn't @paulusmack fix-up some of the signal handling in pppd in the last release (or was that chat.c?). You reported seeing the stats being...

For testing, you could remove the lines in pppd/main.c that say things like "SIGNAL(SIGABRT, bad_signal);". I don't want to do that in the production version.

What version of PPP are you using, and are you using a distro version, and have you made any local modifications? The reason I'm asking is that this is the...

Yes, Jaco Kroon wanted me to be as vague as possible in the public commit, so I'll remove it along with other old stuff and not mention the bug in...