On Mon Aug 30 07:41:18 EDT 2010, [email protected] wrote:
> On Mon Aug 30 07:22:48 EDT 2010, [email protected] wrote:
> > > 1. <insert> in acme on a string like /usr/quanstro/src/
> > > inserts another / for each keypress.
> >
> > I don't know what this means.
> > Acme has no <insert>.
> > I haven't seen this.
>
> oh, sure it does.
>
> /mnt/term/usr/quanstro/plan9/src/cmd/acme/text.c:792,799
> switch(r){
> case 0x06: /* ^F: complete */
> case Kins:
> rp = textcomplete(t);
> if(rp == nil)
> return;
> nr = runestrlen(rp);
> break; /* fall through to normal insertion case */
>
> oddly¸ i see that the arm port has exactly the same bug.
sorry, i take it back. drawterm also has the same bug.
there are several timing bugs, but often what i see
is the +Errors window opening up with just a /
in the body. (but sometimes the / is inserted in
the original space.) Undo in +Errors yields the normal
completion list. texttype() really does get
Kins then '/'.
it turns out that x helpfully yields this if i hold down
the insert key
<insert>//////////////////////////////////////////////
x has a brand new bug.
modern linux systems can be a good example of how better can be
the opposite of working.
- erik
