[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-readline] Pasting large amounts of text into readline-enabled p
From: |
Margarita Manterola |
Subject: |
Re: [Bug-readline] Pasting large amounts of text into readline-enabled programs truncates parts of the lines of the text being pasted |
Date: |
Tue, 23 Jul 2013 12:02:27 +0200 |
Hi,
On Mon, Jul 22, 2013 at 8:53 PM, Margarita Manterola
<address@hidden> wrote:
> that is completely independent of the kernel
I'm sorry for this red-herring yesterday. As I mentioned before,
there was a change in rlwrap's implementation, which tainted my tests.
I'm now testing with a simplified version of the program originally
attached to the report. I'm attaching here the simplified version.
I tested this in a real machines that I had access to:
On Debian Lenny: the program works correctly, kernel version 2.6.26-2-686
On Debian Squeeze: works correctly with 2.6.32-5-amd64, both with
readline5 and readline6.
On Ubuntu Lucid: works correctly with 2.6.38.8 (x86_64) with readline6.1
On another Ubuntu Lucid: fails with 3.0.0-32 (x86_64) with both
readline5 and readline6.1
The program also fails when run on 3.2 (Debian Wheeze), 3.5 (Ubuntu
Quantal), 3.9 (Debian Sid).
I'm thinking of doing a git bisect between 2.6.38 and 3.0, but that
doesn't sound very appealing :-/
Also, regarding what Chet had said, I can confirm that commenting out
one of the calls that set/unset the terminal modes
(rl_prep_term_function or rl_deprep_term_function) "fixes" this (not a
real fix since this breaks the terminal, but it shows that the problem
is in what those calls do; it's interesting that commenting *either*
of them makes this work).
I'm guessing some ioctl call changed its behavior between 2.6.38 and
3.0. I'm unsure if this is readline's fault for using ioctl in a
wrong way, or the kernel's fault for breaking it. I guess it depends
on which call it is and how readline is using it.
--
Besos,
Marga
minirl.c
Description: Text Data