A very simple patch stagnated in the PR database
Jukka A. Ukkonen
jau at iki.fi
Mon Aug 25 08:55:48 UTC 2008
Greetings from Finland,
Are there any committers willing to handle the following
PR and patch?
The issue is very straight forward, but still it has been
left in open state in the PR database for more than a year
now.
http://www.freebsd.org/cgi/query-pr.cgi?pr=standards/116221
This was classified in a bit ad hoc manner as a standards
issue, because there is no "compatibility" category to put
it in, and kern and standards happen to be the best matches
to choose from. I guess I selected standards only because
WNOWAIT is a SUS issue anyhow.
This is very much a "low hanging fruit" type of an issue.
This PR and patch add the flag WNOWAIT to the wait*() system
calls.
The idea of the flag is to allow independent library modules
and threads to avoid collecting the exit status from child
processes they did not fork themselves.
Instead of doing plain...
pid = wait4 (-1, &status, 0, &res);
it will become possible to do...
pid = wait4 (-1, &status, WNOWAIT, &res);
and only if pid is known to the calling module continue ...
pid = wait4 (pid, &status, 0, &res);
reaping only the exit status of a known child.
E.g. SunOS/Solaris already has the WNOWAIT flag.
I have had my own kernel sources updated from the latest
RELENG / STABLE branch automatically on daily basis for
several years and my own patches automatically rolled on
top of that. The WNOWAIT patch has been one of those
automatic patches and a steady part of my system for at
least 18 months now with no ill effects what so ever.
I assume there is now someone thinking "Right, the waitid()
call is also a SUS issue. Why not tackle them both at the
same time?"
I was originally planning to do both waitid() and WNOWAIT,
but there was something related to filling the siginfo struct
for waitid() which was not quite obvious when I added WNOWAIT.
So, I left waitid() out waiting for a moment of inspiration
but added the new flag because the latter was useful on its
own.
It is sad to see these low hanging fruit being forgotten and
left lingering around in the PR database for all ages while
simply merging them and extending the functionality for the
benefit of the whole community would take almost no time at
all.
An outright decline would be better than just leaving simple
patches hanging on the rope. So, it would be nice to find
someone willing to inspect the PR and get it closed one way
or the other.
Cheers,
// jau
.--- ..- -.- -.- .- .- .-.-.- ..- -.- -.- --- -. . -.
/ Jukka A. Ukkonen, Oxit Ltd, Finland
/__ M.Sc. (sw-eng & cs) (Phone) +358-500-606-671
/ Internet: Jukka.Ukkonen(a)Oxit.Fi (Home) +358-9-6215-280
/ Internet: jau(a)iki.fi
v
.--- .- ..- ...-.- .. -.- .. .-.-.- ..-. ..
+ + + + My opinions are mine and mine alone, not my employers. + + + +
More information about the freebsd-standards
mailing list