bug#46502: 28.0.50; [feature/native-comp] (d3a399dd) native-comp bootstrap failure

[Pip Cet]

On Sat, Feb 20, 2021 at 8:56 AM Eli Zaretskii <eliz@gnu.org> wrote:
> > From: Pip Cet <pipcet@gmail.com>
> > Date: Sat, 20 Feb 2021 04:35:39 +0000
> > Cc: akrl@sdf.org, mwd@md5i.com, 46502@debbugs.gnu.org
> >
> > Okay. Here's what I'm seeing (master branch, interrupted after
> > macroexp.elc is compiled).
>
> It's hard to see the differences, since your MUA seemed to have folded
> long lines.

Oh no! Sorry about that.

> Can you send the diffs as an attachment?

I'm sending the complete diff, even though it's a bit large... The
path differences are obviously okay, and I'm not going to complain
about differences in gensym numbering, but byte-optimize-lapcode, to
pick an example at random, differs in the bytecode.

> And when exactly did you interrupt the second run?  The first things
> that get compiled are these:
>
>   COMPILE_FIRST = \
>           $(lisp)/emacs-lisp/macroexp.elc \

HERE

>           $(lisp)/emacs-lisp/cconv.elc    \
>           $(lisp)/emacs-lisp/byte-opt.elc \
>           $(lisp)/emacs-lisp/bytecomp.elc \
>           $(lisp)/emacs-lisp/autoload.elc
>
> Which one of these was being compiled when you pressed Ctrl-C?

cconv.el, after macroexp.el had been compiled successfully.

> > > In general, Make itself will delete any target files it knows about
> > > that were not fully built at the time of SIGINT.  Maybe we don't tell
> > > Make enough about the files native-comp produces?
> >
> > So we should delete the early-compilation .elcs if we end up deleting
> > bootstrap-emacs.pdmp, too?
>
> No, I don't think so.  (What is the relation between
> bootstrap-emacs.pdmp and the produced *.elc files?)

A circular one. bootstrap-emacs.pdmp includes status derived from the
.elc files if those are present before the dump, and it is used to
generate the .elc files otherwise.

> If the
> compilation of these *.elc files has finished, I see no reason to
> delete them, because they should be correct and usable.

Correct and usable, yes. Identical to what they would otherwise have been, no.

bug46502.diff.gz
Description: application/gzip