Re: groff 1.25.0.rc2 on Alpine Linux

Hi Bruno,

At 2026-08-03T17:30:43+0200, Bruno Haible wrote:
> On Alpine Linux 3.24, I see two test failures:
[snip]
> Indeed, line 56 of the generated 'neqn' script looks like this:
> 
>   exec @g@eqn -T ascii "$@"
> 
> The reason is that the 'sed' program on this platform (from BusyBox)
> ignores '-e' options when a '-f' option is present, regardless whether
> the '-e' options come before or after the '-f' option.

Filed as <https://savannah.gnu.org/bugs/?68601>.

How perverse!

POSIX Issue 8 documents groff's usage to date as supported by conforming
implementations.

Even POSIX Issue 4 from 1994 suggests that the invocation I used would
be supported by conforming implementations.

I don't think Busybox had been born yet back then.

Grumble grumble.

> Here's a proposed fix.

Thank you!  Will apply.

> Verified by manually modifying the Makefile accordingly. (I can't
> regenerate a snapshot tarball for testing, because of the missing
> epsf.texi problem
> <https://savannah.gnu.org/bugs/?68566>).

I didn't realize the problem was that crippling.  I'll raise its
priority...especially since I'm making Git users use maintainer mode.
:-O

Regards,
Branden

-----BEGIN PGP SIGNATURE-----
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=vQ9n
-----END PGP SIGNATURE-----