I think we've covered this bit by now. On Wed, Dec 7, 2011 at 3:32 PM, Antonios Motakis <a.motakis at virtualopensystems.com> wrote: > On 12/07/2011 08:52 PM, Christoffer Dall wrote: >> >> >> In which case it would not work anyhow - the boot string is embedded >> inside the binary and a re-compile cannot be avoided. > > Yes I know, my point is that the user is not exposed to the fact that his IP > address is being used. Before, the developer did know that his IP is used in > the binary, because he put it in there himself, so he knows to do the right > thing. > > >> I think we are not talking about casual users using boot-wrapper in >> any case. It's just a question of making it easy for developers to run >> the whole setup. > > Agreed, my point is that if it is being done by a script that is called > without the user being aware of it, then even a developer might miss the > fact that the output is specific to the host IP address. > > Setting it in a config file or Makefile makes it pretty clear, even if by > default we but with backticks a command returning the current IP; however > the way it is now, where it is being called transparently from Make, it can > be confusing, because even developers are usually not used to this kind of > sorcery. > > Actually, it is worse than that; the current behavior doesn't just require a > recompile, a make clean is required as well. Make doesn't detect that the > host IP has changed. These things might be obvious to us, but if you have > not read the code they are very far from that. > > >> >> Personally, I just want to go to the boot-wrapper directory, do 'make' >> and be on my merry way. Not copy output from ifconfig manually and >> stuff if I can avoid it. > > I agree with that, I just wanted to point out the potential for confusion.
