On Sat, Feb 07, 2026 at 05:51:02PM -0600, Tim Chase wrote:
> On 2006-09-14 03:23, Bob Beck wrote:
> > When I'm a sysadmin and type vi, I want vi with all it's ususal
> > idiosyncracies so that it's basically the same no matter what system
>
> Idiosyncracies aside, I'd still like to see it get bug fixes, such as
> this one that regularly bites me:
>
> $ yes hello | head -50 | fmt -100 > long_lines.txt
> $ vi long_lines.txt
> :g/^/.!fmt
>
> results in
>
> +=+=+=+=+=+=+=+
> Illegal address: only 6 lines in the file.
> Error: unable to retrieve line 7; 4 lines added; 2 lines deleted
>
> instead of re-fmt(1)ing each line individually. Another case involves
> duplicating each line in the file:
>
> $ vi
> :r !jot 100
> :g/^/t.
>
> which produces:
>
> +=+=+=+=+=+=+=+
> Illegal address: only 150 lines in the file.
> Error: unable to retrieve line 151; 50 lines added
>
> This also ends up duplicating *every other* line, instead of duplicating
> *every* line.
>
> -tkc
>
> PS: FWIW, the problem exists in FreeBSD's vi/nvi too:
>
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=270048
>
> which I reported there at the same time I reported this on bugs@
>
> https://marc.info/?l=openbsd-bugs&m=167830927129031&w=3
>
> but never heard anything on either front.
> --
>
Index: ex/ex_global.c
===================================================================
RCS file: /cvs/src/usr.bin/vi/ex/ex_global.c,v
diff -u -p -u -p -r1.17 ex_global.c
--- ex/ex_global.c 27 May 2016 09:18:12 -0000 1.17
+++ ex/ex_global.c 8 Feb 2026 09:48:32 -0000
@@ -269,12 +269,12 @@ ex_g_insdel(SCR *sp, lnop_t op, recno_t
/* If range less than the line, ignore it. */
if (rp->stop < lno)
continue;
-
+
/*
- * If range greater than the line, decrement or
- * increment the range.
+ * If range greater than or equal to the line,
+ * decrement or increment the range.
*/
- if (rp->start > lno) {
+ if (rp->start >= lno) {
if (op == LINE_DELETE) {
--rp->start;
--rp->stop;
--
Walter
