search for: urgghh

On 4/15/20 11:16 AM, Richard W.M. Jones wrote: In the subject, you describe $nbdkit_safe_stdio, but in the patch body... > --- > plugins/eval/nbdkit-eval-plugin.pod | 11 +++-------- > plugins/sh/nbdkit-sh-plugin.pod | 18 +++++++++++++++++- > plugins/sh/call.c | 8 ++++++-- > tests/test-single-sh.sh | 4 ++++ > 4 files changed, 30

...est-single-sh.sh | 4 ++++ > > 4 files changed, 30 insertions(+), 11 deletions(-) > > > > The patch does not work as intended as currently written: when we > invoke /path/to/script config $key $value, we have already set up > stdin/stdout to our own pipes [1]. Urgghh ... very true. I think patches 1-8 are still worthwhile though :-) And with the same mechanism we can pass other environment variables in if we think of any in future ($nbdkit_peer_name?). > For .config and > .config_complete, reading will always see EOF (no other callback > needs to in...

...4 ++++ >>> 4 files changed, 30 insertions(+), 11 deletions(-) >>> >> >> The patch does not work as intended as currently written: when we >> invoke /path/to/script config $key $value, we have already set up >> stdin/stdout to our own pipes [1]. > > Urgghh ... very true. > > I think patches 1-8 are still worthwhile though :-) And with the same > mechanism we can pass other environment variables in if we think of > any in future ($nbdkit_peer_name?). Indeed, and I plan to review them (a quick glance made it seem like they are nice, but...