On 2023-06-14 Andreas Metzler <ametz...@bebt.de> wrote:
> On 2023-06-14 Andreas Beckmann <a...@debian.org> wrote:
[...]
> > after building gnutls28/experimental with pbuilder, there are a lot of
> > gnutls-serv and a few openssl s_server processes running, preventing
> > pbuilder to cleanup the chroot.
> This was caused by the changes related to fixing #1031553 (supporting
> either datefudge or faketime.
The reasons for breakage are twofold:
when using a wrapper shell-function instead of an executable like this:
wrapme() {
foo "$@"
}
This kind of code breaks
wrapme server &
PID=$!
...
kill $PID
since PID will contain the process ID of the currently running script
instead of "foo server".
The second one is a similar one. While
datefudge ... server
gives you one process - the server one (datefudge is a shell script
which sets some environment variables and exec-s the a binary.) OTOH
faketime ... server
starts a faketime process and a forked "server" as child process and
killing faketime does not kill the child process (Reported as
https://github.com/wolfcw/libfaketime/issues/428 ). This also breaks the
idiom mentioned above.
cu Andreas
--
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
