Re: autoconf hang

[Paul Eggert]

Troy Heber <address@hidden> writes:

> ...
> 23052 execve("/usr/bin/m4", ["/usr/bin/m4", "--nesting-limit=1024"], [/* 43 
> vars */]) = 0
> 23052 uname({sys="Linux", node="me", ...}) = 0
> 23052 brk(0)                            = 0x805d000
> ...

This is the normal way that processes start up, and it looks OK.  There
were lots of other execve calls like this in your trace.

> ...
> 23056 execve("/bin/sh", ["sh", "-c", "/usr/bin/m4 --nesting-limit=1024"...], 
> [/* 42 vars */]) = 0
> 23056 --- SIGINT (Interrupt) @ 0 (0) ---
> ...

This is the process that was hanging.  It loos like it was hanging
trying to do the execve -- that is, the subsidiary shell did not
take control for some reason.

One possibility is that your operating system ran into some resource
limitation issues (number of simultaneously executing programs,
perhaps?) that causes it to hang in this siutation.

Another possibility is that /bin/sh got into an infinite loop before
it invoked uname.  (Perhaps the dynamic linker looped?).

I can't think of any other possibilities.  But perhaps you can investigate
further, e.g., by attaching a debugger to the "hung" process.