| Summary: | sigaction(2) man page is misleading | ||
|---|---|---|---|
| Product: | Documentation | Reporter: | Ronald F. Guilmette <rfg> |
| Component: | Books & Articles | Assignee: | freebsd-doc (Nobody) <doc> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | ||
| Priority: | Normal | ||
| Version: | Latest | ||
| Hardware: | Any | ||
| OS: | Any | ||
State Changed From-To: open->analyzed I committed a shortterm fix which at least shows the up-to-date source instead of the old one. I need to delve in my POSIX docs before I can change the wording. But at least it cannot be understand at this point. Responsible Changed From-To: freebsd-doc->asmodai I need to get some docs and look this over. Responsible Changed From-To: asmodai->freebsd-doc Move back to the doc pool. Using FreeBSD 5.0-Current, the manpage I have seems to have been updated since 4.1 and I was wondering if this change was done in 4.4 Release? If so, should this be closed? Tom (darklogik) Rhodes http://www.Pittgoth.com The Dark Portal Project http://www.FreeBSD.org The Power To Serve This pr is old, (look at the date) and I wanted to make sure that the current man page was not bothered. 4.4 4.5 and 5.0-CURRENT man pages have already been updated, I don't know how far back as I did not do version checking. With this in effect I highly doubt this pr needs to remain open, any comments? Tom Rhodes www.Pittgoth.com www.FreeBSD.org State Changed From-To: analyzed->closed Man page updated in both -current and 4.4, so this can be closed. Thanks for the report. |
The sigaction(2) man page is misleading with regards to the current (new) definition of the `struct sigaction' type. That type is now defined as: struct sigaction { union { void (*__sa_handler) __P((int)); void (*__sa_sigaction) __P((int, struct __siginfo *, void *)); } __sigaction_u; /* signal handler */ int sa_flags; /* see signal options below */ sigset_t sa_mask; /* signal mask to apply */ }; But the definition given in the sigaction(2) man page leads one to believe that the following might be valid C code: struct sigaction thing = { handler, 0, 0 }; but if you do that, with -Wall, gcc complains about a missing set of curly braces. Fix: Just say what the POSIX standard says, i.e. ``The `struct sigaction' type contains at least the following members...'' Just say that on the man page, rather than attempting to actually show (inaccurately) the definition of the sigaction structure. How-To-Repeat: See above.