axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] your patches


From: root
Subject: [Axiom-developer] your patches
Date: Mon, 28 Jul 2003 20:11:11 -0400

Juergen,

I've applied some of your patches but left others out.

(a) The SpadInterpretStream patch is ok. It should be a number.
(b) The funPLUSform ... patch is ok.
(c) The LED/NUD table patch is suspect. Why do you think we need to
    change the LED/NUD value for +-> and why is 112 the right value?
(d) data patch. I couldn't prove that data was ALWAYS unbound so I
    added the line:
    if (BOUNDP 'data) then ...
    It has the same effect as removing it as you did but is safer.
    I commented out the data usage on the initialization line as you did.
(e) the indexList -> indexVars change was not made. It requires more study
    on my part before I'm comfortable with it.
(f) the LET code is unchanged. LET takes 2 values but this is a compiler
    internal form, not a lisp (let form so I have to prove that only
    2 of the three arguments are ever used.
(g) the suggested autoLoad change will break autoLoad. If the autoLoad
    doesn't work then the whole build process is broken and the infinite
    load loop isn't the issue. This change was not applied.
(h) TRUE -> true, NIL -> false was applied

In general I need to PROVE that a change to the code won't break
anything else before I apply it. That involves following all possible
code paths (as in the SpadInterpretStream case). This is painful but
necessary to ensure that we don't break things. Sometimes the interactions
are very subtle. As a side-effect I ended up writing a simple explanation
of the top level loop code so, while painful, it was useful as it forced
me to document.

Tim
address@hidden
address@hidden




reply via email to

[Prev in Thread] Current Thread [Next in Thread]