Hi Bruno,

At 2026-02-21T01:01:59+0100, Bruno Haible wrote:
> Testing this prerelease on those platforms where groff 1.24.0.rc2 had
> problems:
> 
> * On CentOS 7 and Alpine Linux 3.18, the build now succeeds and all
>   tests pass — even without configure option --without-urw-fonts. In
>   other words, the unavailability of the URW fonts was detected at
>   build time.  Perfect!

Great news--thanks!

> * FreeBSD 14.0 and 15.0, NetBSD 10.0, OpenBSD 7.8 there is 1 test
>   failure:
>   FAIL: src/preproc/tbl/tests/boxed-TH-tables-work-on-utf8-device.sh
>   Find attached the log files.

Thank you.  All of these are same (presumably the same implementation of
od(1), or genetically related ones, are in use), and they appear to be
spurious.

My regex is too strict:

---snip---
echo "checking that right edge of table after heading looks like -|" >&2 
echo "$output" | grep -q '0000040 .*  *e2  *94  *a4  *0.$' || wail
---end snip---

I failed to anticipate that these "od"s would pad the output line with
spaces to the 73rd column.

Regards,
Branden

Attachment: signature.asc
Description: PGP signature

Reply via email to