Inetd.c sigpause() --> __xpg_sigpause() not in uClibc
John Z. Bohach
jzb at aexorsyst.com
Wed Nov 9 04:20:52 UTC 2005
> I'm not against rewrites. But complete rewrites masquerading as a bugfix
> for something that has a 2-line patch, and thus go in unannounced?
>
> I never noticed this before because inetd is something I simply don't
> use...
>
> Rob
Well, I was afraid it might be something like this, thus I didn't want to
simply throw another patch at ya...
Thanks for the analysis. I've done a bit more research on the '__xpg_'
prefix, and it seems that sigpause() is an XPG/4 or XPG4-Unix extension to
the signals.c in the C library.
Its likely that the true culprit here is uClibc not providing __xpg_sigpause
and only providing __sigpause (this IS a difference from glibc), but I'm not
sure if Erik wants to add a deprecated prefix into uClibc. This has
something to do with the -D_XOPEN_SOURCE preprocessor flag or some other
equally standards-related flag.
Anyway, a quickie fix is simply to revert the code around the sigpause() back
to 11010, and in looking at that old code, I don't see anything obviously
wrong with it, but then again I've not done a thorough code analysis on that
portion. If you need me to do so, and actually send a patch back, I don't
mind at all, since that's probably what I'll do anyway on my local copy of
busybox just to get past this. But if I really feel motivated, I might add
the __xpg_sigpause to uClibc and send that as a patch to the uClibc project.
It somewhat depends on what you and others think is the best approach...
--john
--
### Any similarity between my views and the truth is completely ###
### coincidental, except that they are endorsed by NO ONE ###
More information about the busybox
mailing list