From mboxrd@z Thu Jan  1 00:00:00 1970
From: erik quanstrom <quanstro@quanstro.net>
Date: Mon, 30 Aug 2010 08:39:51 -0400
To: quanstro@quanstro.net, rsc@swtch.com, 9fans@9fans.net
Subject: Re: [9fans] P9p on Fedora 12
Message-ID: <1c5d1df7f94e7d87465fb1539a28fd6a@plug.quanstro.net>
In-Reply-To: <4bd7aaa3effec6d1152b864f1c28bf73@plug.quanstro.net>
References: <161c27f3-93af-4f06-86de-7d1c7e473a4b@u20g2000yqu.googlegroups.com>
	<AANLkTimVDVkJvV5c6nYf_-VQSCyhEg7gu1bpBEPN6qgr@mail.gmail.com>
	<75395937b512b64fc9892f1401515e04@plug.quanstro.net>
	<AANLkTimPR51g2Zh9CK76rFOH0SVou_0frOXoky9o_1hC@mail.gmail.com>
	<4bd7aaa3effec6d1152b864f1c28bf73@plug.quanstro.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8bit
Topicbox-Message-UUID: 4ac10a38-ead6-11e9-9d60-3106f5b1d025

On Mon Aug 30 07:41:18 EDT 2010, quanstro@quanstro.net wrote:
> On Mon Aug 30 07:22:48 EDT 2010, rsc@swtch.com 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