Update of bug #67838 (group groff):
Status: Need Info => In Progress
Assigned to: None => gbranden
_______________________________________________________
Follow-up Comment #12:
Hi Alexis,
Your patch (file #58069) did not apply correctly because:
1. It had no context. "groff.texi.in" is often a fast-changing file in my
working copy, and cases like that are where context (including "universal")
diffs really shine.
2. Your patch recoded the file, on the lines that it changed, from Latin-1 to
UTF-8. Don't do that. ;-)
But I see the thrust of your patch. Some version of Texinfo appears to have
relieved the document author of the necessity of hand-maintaining the topology
of the node tree in two different ways. If Texinfo 5.0 can't handle that,
it's worth bumping our minimum version requirement to the first that can (if
it's not ridiculously recent).
The foregoing necessity is one of the most hostile anti-features any piece of
GNU software came up with. I still haven't figured out the structure, which
is why I made so many mistakes hand-maintaining these data. It *seems* like
the idea is to perform a breadth-first traversal of the document in the
forward ("Next") direction, but *depth*-first in the backward ("Prev")
direction.
If true, I find that psychotically unintuitive.
I suppose some piece of Elisp automatically produced Texinfo @node commands
that recapitulated the node topology otherwise inferable from @node sequencing
in the input and the `@menu`/`@end menu` blocks, and the general view among
parisioners of the Church of Emacs was that all of the infidels could just
f*** off.
...which explains why Texinfo succeeded brilliantly at its mission of
eradicating man pages forever. :-|
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?67838>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
