translating in-source builds, next round (was Re: gcc/gcc ChangeLog doc/install.texi)
Tue Jun 18 16:05:00 GMT 2002
> build a broad consensus that this is the wrong way to go, and I will
> follow that consensus.
As a build machinery maintainer, I shouldn't have to build a
consensus. A patch was proposed. I rejected it. You approved it. I
can't be a maintainer if this is how things are going to work.
> Phil's patch can't possibly prevent you from doing your work. (You can
> always remove it in your tree until your work is done.)
I *did* ask you to remove it from mainline, which is where we're doing
> > It also glosses over all the work I and others have done to make
> > in-source builds work (they're tested nightly by at least two of us
> > now).
> Is that true for the 3.1 branch, the mainline, or both?
Mainline at least. I don't recall if the patches were applied to the
> > But I see it was approved and checked in anyway, so I guess we're
> > already screwed. Mark, please make sure this gets approved and
> > applied to the binutils/gdb repository, if you can.
> I don't have that authority.
To submit a patch? Of course you do. Whether you can get it approved
or not is unknown. If you can't convince them to approve it, then
perhaps you shouldn't have approved it for gcc either.
> You'll have to take it up with the binutils and GDB maintainers if
> you'd like the patch applied there.
We had this conversation a few months ago, and everyone who voiced an
opinion agreed that the toplevel files should be kept in sync between
gcc and binutils/gdb. I've been keeping them in sync since then,
either by applying patches manually (I have suitable rights on each
side) or by asking the contributor to propose the patch for the other
side and getting approval. Since I don't approve of this patch, I
won't do it manually, so I'm asking you to email the patch to the
binutils/gdb/newlib/cygwin/sid mailing lists and ask for them to also
If you can't get their approval, then I, as build machinery
maintainer, am obliged to ask you to revert it, since the concensus is
that we should keep those files in sync.
More information about the Gcc-patches