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

Re: 4.2

I'm getting confused about exactly what you're after here.  If you
want to say that a compiled file is guaranteed to load correctly only
in the implementation it was originally compiled in, that's OK with me
(although I think it goes without saying since the standard leaves the
format of such files unspecified).  But I still don't understand why
you think it's necessary to require things like the compiled file
aving to be loaded into a "fresh" copy of the Lisp image.  Can you be
more specific about the rationale for doing this?  What particular
advantage does it buy implementors that would justify putting such a
restriction on users?  What kinds of things would the compiler be
assuming about the code that would be invalidated if it were not
loaded into a "fresh" image?  Among other things, this restriction
implies that you can't load more than one compiled file, since the
image would no longer be "fresh" after loading the first one.  I don't
think any existing implementation is currently this restrictive, and I
imagine that any implementation that did place such restrictions would
be dismissed as a toy. 

Like I said before, I feel very uncomfortable with the idea of changes
in content like this being stuck into the standard at the last minute
using only the editorial process, without first having some discussion
and a vote of X3J13 as a whole.  If you want to write up a specific
proposal on this, though, I'm willing to bring it up for consideration
at the meeting later this month, along with the rest of the unresolved
cl-compiler issues.