Re: [Nmh-workers] nmh current development

[Paul Fox]

 > > > though.  I'll even stop whining about putting autoconf-generated in
 > > > CVS (I'll still think that it's a huge mistake, I just won't say
 > > > anything about it :-) ).
 > > 
 > > do you think it's a mistake even if they're commited under
 > > non-standard names?  i agree that they probably shouldn't be committed
 > > under their real names, because that does cause confusion, and can
 > > lead to inadvertent commits.
 > 
 > What's the point of committing them at all?  They're what's produced by
 > the committer's particular version of auto*.  It's like me checking in
 > main.o -- it reflects my compiler chain.

it's a far cry from that.  main.o represents your architecture,
your OS, etc.  committing a configure script is a lot more
similar to committing, uh, processed man pages in addition to
their nroff source, for people that don't have nroff.  the configure
script is widely useful as a product, and having _any_ configure,
even one not 100% correct, is better than having none.

 > Anyone expecting to build from CVS should also be prepared to have
 > suitable auto* tools.  The tar files of source releases contain the
 > auto* output so that's useful for posterity and debugging build bug
 > reports.

okay, okay, okay.  forget about commiting the generated files. 
i'll upgrade my auto* tools.  sheesh.  :-)

and as ken so wisely points out, more regular source releases
would help a lot too.  ;-)

it sounds like a couple of people have some fresh or renewed
energy for making progress.  given that, it would be really
really great if this thread didn't die off like others like it in
the past.  to keep the rest of us from waiting in suspense, are
ken and jon working out a new maintenance agreement of some sort?

paul
=---------------------
 paul fox, address@hidden (arlington, ma, where it's 52.3 degrees)