On Sun, Aug 27, 2000 at 02:45:18PM +0100, Daniel Barlow wrote:
> William Harold Newman <william.newman@...> writes:
> > True enough, that looks useful. I have
> > * renamed LOAD-OBJECT-FILE to LOAD-1-FOREIGN and exported it from SB-EXT;
> > * changed its return value to T (on the theory that it's like CL:LOAD); and
> > * added some documentation based on your example above.
> Cool! Can I make a request, though? Find another name ...
> If I were using apropos to find it, my guesses would be functions
> containing the strings "LOAD" (which, granted, would find it),
> "ALIEN", "OBJECT", or "SHARED". I don't know where this practice of
> calling random things in the ALIEN system "FOREIGN" instead started[*],
> but it only affects a very few external symbols, so if this is an
> opportunity to clean it up, I'd urge that that happens.
> If it were called, say, LOAD-ALIEN-SHARED-OBJECT, it's even obvious
> (obvious to a Unix programmer who's meddled with shared libraries
> before) what format the file it takes as an argument. And that it's
> OBJECT not OBJECTS eliminates the need for the odd-looking "-1-"
I named it LOAD-1-FOREIGN because it's a degenerate case of full-blown
LOAD-FOREIGN, and I expect that someday RUN-PROGRAM will be
implemented and LOAD-FOREIGN can be implemented easily thereafter. I
wanted the names to be consistent, and it didn't occur to me to rename
> OK, I'm picky. Having the facility at all is a much bigger deal - if
> you don't change the name just for me, I'm not going to be too upset.
Well, anyone who's familiar both with SBCL and with CMU CL can tell
that I'm pretty quick to rename things when I think the old names are
unclear or inconsistent or misleading. I'll take a look at the various
public names with ALIEN and FOREIGN in them and think about
> (Alpha port, for the interested, has now reached the point of dying
> somewhere else in cold-init, and changing its behaviour when I insert
> (%primitive print) statements. Anyone know how to usefully interpret the
> addresses in ldb backtraces?)
That sounds promising.
I'm sorry I can't give any useful advice about ldb, though, since I
just used lots and lots of /SHOW0 statements to tell where problems
occurred, and then used gdb to inspect values, until I got to the
point I could use the full debugger.
William Harold Newman <william.newman@...>
PGP key fingerprint 85 CE 1C BA 79 8D 51 8C B9 25 FB EE E0 C3 E5 7C