Thanks for the release.  Rather above and beyond the call of duty.

You wrote:

>  Changes mostly by Nick and Sarathy while I was offline, Chris has also
>  been been busy on the macperl branch.  I still have to read more carefully
>  the p5p archives; I may disagree somewhat with the commonsense.t change
>  (IIRC the idea was that there are certain libraries without which Perl
>  is quite unusable, that's exactly why it was in t/base)

I'll re-iterate that moving commonsense.t to t/lib seems to be a more
natural place, allows the minitest make target to run, and will still
indicate failure if something is awry even if it is moved from t/base to
t/lib.

That said the failure of `make test` is obviously much more spectacular
with commonsense.t in t/base.  Hence perhaps we could leave
t/base/commonsense.t in place (or move it back over since the 11446 kit 
has it in t/lib) and add these minitest instructions to the INSTALL file:

--- install;1	Mon Jul 23 12:54:25 2001
+++ INSTALL	Mon Jul 23 13:00:58 2001
@@ -1464,6 +1464,7 @@
 If you can successfully build miniperl, but the process crashes
 during the building of extensions, you should run
 
+	mv t/base/commonsense.t t/lib
 	make minitest
 
 to test your version of miniperl.
End of possible doc patch.

Hmm let's see, for Windows without unix-like utilities (e.g. %COMPSPEC% 
= cmd.exe) you would need to say either:

+	copy t/base/commonsense.t t/lib
+ 	dmake minitest

or,

+	copy t/base/commonsense.t t/lib
+ 	nmake minitest

For VMS you would need to say:

+	copy [.t.base]commonsense.t [.t.lib]
+ 	mms minitest

etc. which would be too verbose.  At least the unix shell patched version
ought to provide an easily translated clue.  Note that folks most likely
to be running the minitest make target are those more likely to have read
the build instructions anyway, hence an extra setup step for the minitest
target may not actually be asking too much of them.

Peter Prymmer