At 2026-01-18T14:47:33+0100, Ingo Schwarze wrote: > Gavin Smith wrote on Sat, Jan 17, 2026 at 07:08:34PM +0000: > > On Sat, Jan 17, 2026 at 12:30:38PM -0600, G. Branden Robinson wrote: > >> At 2026-01-17T19:13:35+0100, Patrice Dumas wrote: > >>> On Sat, Jan 17, 2026 at 11:40:27AM -0600, G. Branden Robinson > >>> wrote: > > >> Thanks, Patrice! Looks like I could have put this change in for > >> groff 1.23.0, which bumped our Texinfo dependency from 4.8 to 5.0. [...] > > > It's a lot older than Texinfo 5.0, as far as i gather. > [...] > > This comment is unchanged from Texinfo 3.1, from 1993. (I haven't > > tried building any of these old releases to check how automatic node > > pointers work, though.) > > > > I don't know if many of the releases in the Texinfo 3.* series are > > still available anywhere. On ftp.gnu.org there is only 3.0, 3.1, > > 3.9 and 3.12 (I remember some of those were not original tar > > distributions and were sources I salvaged from other places on the > > Internet.) > > For what it's worth, OpenBSD-current ist still shipping with this > very ancient texinfo in the base system: > > $ uname -a > OpenBSD isnote.usta.de 7.8 GENERIC.MP#209 amd64 > $ sysctl kern.version > kern.version=OpenBSD 7.8-current (GENERIC.MP) #209: > Fri Jan 16 00:14:51 MST 2026 > [email protected]:/usr/src/sys/arch/amd64/compile/GENERIC.MP > $ makeinfo --version > makeinfo (GNU texinfo) 4.8 > Copyright (C) 2004 Free Software Foundation, Inc. > > We are unwilling to update to a newer version. If i recall correctly, > the main reason is licensing, because newer TexInfo is no longer > available under the slightly more free GPLv2 and now comes with the > even more encumbered GPLv3.
I remember now why I bumped the version dependency from 4.8 to 5.0. It wasn't to foist the scary, scary GPLv3 on people. It was because groff's Texinfo manual needed its `@codequoteundirected` command. groff, being a typesetting system, distinguishes many varieties of quotation mark, and it was strongly confusing to the user to have the wrong ones show up in examples and discussion of quotation mark symbols. One could infer the foregoing from reading the full commit to which I linked earlier in the thread, but the following commented code goes directly to the issue. https://cgit.git.savannah.gnu.org/cgit/groff.git/tree/m4/groff.m4?h=1.23.0#n105 Regards, Branden
signature.asc
Description: PGP signature
