Re: fork not working?

The line "p Process.fork do" does not behave as you expect because do
has a low precedence. This makes ruby behave as if you had written:

p(Process.fork) do
  `touch /tmp/bar`
end

which makes a new process without the associated block, when what you meant was:

p(Process.fork do
  `touch /tmp/bar`
end) 

Yuck and yuck, but the script returns, and you don't see an error
code. You could use braces instead of do...end or do something like

id = Process.fork do <code> end
puts id

Though having said all that, I must ask why you want to fork a new
ruby interpreter from within Apache. :)

-mike

On Sun, 09 Jan 2005 20:51:29 -0500, Colin Steele <[email protected]> wrote:
> Hi, all.  I'm running the following mod_ruby script:
> 
> p Apache.server_version
> p ENV["MOD_RUBY"]
> 
> `touch /tmp/foo`
> 
> p Process.fork do
>   `touch /tmp/bar`
> end
> 
> sleep 10
> puts `ls -l /tmp/foo`
> puts `ls -l /tmp/bar`
> exit
> 
> Which produces the following output in my browser:
> 
> "Apache/2.0.51 (Fedora)"
> "mod_ruby/1.2.4"
> 4474
> -rw-r--r--  1 apache apache 0 Jan  9 20:44 /tmp/foo
> 
> So I'm left wondering, where'd my forked child with pid 4474 go?  What's it doing?  How come it didn't touch /tmp/bar?
> 
> Is this a bug or a feature?  I grepped around in the mod_ruby source and couldn't find anything obviously related to Process.fork... any help would be greatly appreciated.
> 
> --Colin Steele
> [email protected]
> 
>