On Fri, 15 Oct 2004, David Berry wrote:
> > Well, I don't quite believe it, but a clean checkout followed by:
> >
> > ./bootstrap
> > make configure-deps
> > ./configure -C
> > make world
> >
> > under RH9 has just ran to completion, without human intervention. Must
> > have done something wrong, but perhaps we should put the champagne on ice
> > and warm up the fat lady...
> >
> > Anyone else seeing similar successes?
>
> So what did you do about the htx dependency issue? The Makefile.dependencies
> in my recent clean checkout (this morning) has nothing depending on htx,
> and the ALL_TARGETS list in Makefile has htx after echomop and ast, both
> of which use htx. So when (and why) did htx get built in your sucessful
> make world?
It's part of ALL_TARGETS, so got built at some point, but you're right it
did fail, but quietly. If I look at the ECHOMOP section in my log it says:
/bin/sh: line 131: hlink: command not found
at the very end, but then continues.
Grepping for hlink I cannot see any mention of it in the AST logs (so I
guess it's not currently used, appears in some script buildhyperdocs, but
no one seems to really use that). Similar story for KAPLIBS, but HTX had
been built by then, so it could have been used.
> I'm currently doing a make world on RH9 from a clean(ish) checkout done
> earlier this morning. Prior to doing "make world" I remade
> Makefile.dependencies to include htx as dependencies of ast, kaplibs and
> echomop. No problems so far, but it's only got as far as libraries/cat.
Sounds like the final nail.
Peter.
|