Subject: Re: Bug in vortex?
From: Craig Chambers (chambers@cs.washington.edu)
Date: Thu Apr 25 2002 - 14:33:31 PDT
That's my bad. I replaced the deps boolean option with the constant true, but I
did this from my whirlwind tree. I forgot to also do this constant-folding from
the vortex source tree as well.
I just checked in the fixed precedence.cecil file. I'm now recompiling vortex,
and then I'll test the new executable, but I think this change will work for
you.
-- Craig
Keunwoo Lee wrote:
>
> While trying to prepare a bootstrapping RPM distribution of Vortex, I
> attempted to build the Cecil interpreter on Linux using a freshly built
> Linux Vortex. Build dies with the following error in boolean.cecil.c:
>
> ++ compiling assertion_failed(@any):void..
> ** Fatal error: message "deps" with 1 arg not understood
> Current stack frame:
> # 0 eval (_anon_closure: <closure>) in form_msg_expr_in,
> precedence.cecil:213
> Locals: right: <anon/send_expr/: 0xa46af81>,
> left: <anon/string_literal_expr/: 0xa453229>
>
> The relevant chunk of code is
>
> if(relationship = "below" |
> { relationship = "same" & { group.associativity = right_assoc } }, {
> -- New operator binds tighter than previous operator. Continue
> -- with next op
> if(options.deps, {
> let d:precedence_dependence :=
> new_precedence_dependence(op, op1, right_assoc);
> d.add(app.current_dep);
> });
> eval(exit);
> }, {
>
> options.deps is the offending call. options is not a local variable. I
> conclude that it is a reference to the global object 'options', in
> options-misc.cecil, which appears to have no method or field named deps.
>
> grep -e '\(field\|method\) deps' *.cecil
>
> reveals
>
> has-nlrs.cecil: var field deps(@:has_nlr_info):nlr_dependence := not_defined;
> ipca-class-set.cecil: field deps(@:monitored_pure_type[`CSDep]):m_set[CSDep] :=
> ipca-class-set.cecil: method deps(@:pure_type):collection[CSDep] { [] }
> method-table.cecil: field deps(mte@:method_table_entry):method_dependence :=
>
> None of these dispatch on options or one of its children. I conclude that
> 'options.deps' is a bug. Can we fix this?
>
> ~k
>
> _______________________________________________
> Cecil mailing list
> Cecil@cs.washington.edu
> http://majordomo.cs.washington.edu/mailman/listinfo/cecil
_______________________________________________
Cecil mailing list
Cecil@cs.washington.edu
http://majordomo.cs.washington.edu/mailman/listinfo/cecil
This archive was generated by hypermail 2b25 : Thu Apr 25 2002 - 14:34:08 PDT