[Mono-dev] [PATCH] Incorrect signal handling for Sys V signal handler

[Miguel de Icaza]

> I was thinking about using a helper function that does "the right thing"[tm], 
> using sigaction if available, with a fallback to sigset and signal. A unified 
> function that may be used in other parts of the code is probably not 
> possible, since sigaction allows a number of flags that influence the signal 
> disposition and signal handling. I'm not sure whether a default set of flags 
> works for all use cases (e.g. syscall restarting yes/no, alternate signal 
> stack). Since signals are modified in a number of source files, a central 
> place to handle signals is probably the best way to go.

Perhaps we need to re-evaluate the use of signal there in the first
place, and use sigaction consistently.

That is what I was alluding to before.