I'm guilty of not telling the whole story when I reported this problem. I
hate people who do that.
There IS a .git directory. I didn't mention it because "it couldn't
possibly affect anything". Just after untarring, I did a "git init",
because I wanted to capture all changes that I make to the source code,
starting with editing config.make. I never dreamed that the actions of the
make file would depend on whether or not a .git directory existed.
My apologies, and thanks for looking at this.
On Tue, Oct 18, 2016 at 2:49 AM, Peter Bex <pe...@more-magic.net> wrote:
> On Tue, Oct 18, 2016 at 12:49:48AM -0400, Terry Phelps wrote:
> > Well, here's some more details, but the timestamps look okay to me.
> OK, so that's ruled out.
> > [...]
> > Near the end of the above, it says:
> > Prerequisite `buildid' is newer than target `build-version.c'.
> > Prerequisite `buildtag.h' is newer than target `build-version.c'.
> > And those two files are indeed newer than the target, so the problem
> > that build-version.scm is newer, I think. It is that buildid and
> > are prereqs, and ARE newer:
> > $ ls -l buildid buildtag.h
> > -rw-r--r--. 1 tgphelps tgphelps 8 Oct 17 14:36 buildid
> > -rw-r--r--. 1 tgphelps tgphelps 76 Oct 18 00:15 buildtag.h
> > I don't know enough about the mechanics of building chicken to debug this
> > any further, so I'll stop here.
> It looks like identify-me.sh is creating these files. It should avoid
> doing that if there's no .git dir and there is a manual-html dir.
> Can you check that the manual-html exists and there's no .git dir in the
> build root?
Chicken-users mailing list