On Sat, Jan 4, 2014 at 12:00 PM, Robert P. Goldman <[email protected]> wrote:
> Yes, I want to disable this on platforms where it is known to fail.
>
> Unfortunately, the backtrace does not make it at all easy to tell where the 
> failure has occurred, so the poor user will have no way to tell that the 
> failure is not his or her fault but the fault of ASDF and the implementation.
>
> On ABCL the bundle files (jar files, IIRC), for some reason, are simply 
> missing files that should be present. On ECL, something goes pear-shaped deep 
> in the interaction with C. Neither test yields an error I can make heads or 
> tails out of.
>
> I was thinking of issuing some form of warning that indicates that here be 
> dragons, and the user is proceeding at his/her own risk. That isn't quite 
> full disablement, but it will keep people from wasting their time in a futile 
> way if it bombs out.
>
What about a cerror? That way, the user can still access the
functionality if he's willing to debug it.

—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org
War does not determine who is right — only who is left. —  Bertrand Russell

Reply via email to