[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190220203309.fdxhriqd6r5tacyt@angband.pl>
Date: Wed, 20 Feb 2019 21:33:09 +0100
From: Adam Borowski <kilobyte@...band.pl>
To: Juerg Haefliger <juerg.haefliger@...onical.com>
Cc: Steven Rostedt <rostedt@...dmis.org>, linux-kernel@...r.kernel.org,
mingo@...hat.com, Masami Hiramatsu <mhiramat@...nel.org>,
Shuah Khan <shuahkhan@...il.com>
Subject: Re: [PATCH] selftests/ftrace: Make the coloring POSIX compliant
On Wed, Feb 20, 2019 at 09:20:20PM +0100, Juerg Haefliger wrote:
> On Wed, 20 Feb 2019 14:49:34 -0500
> Steven Rostedt <rostedt@...dmis.org> wrote:
>
> > On Wed, 20 Feb 2019 17:13:33 +0100
> > Juerg Haefliger <juerg.haefliger@...onical.com> wrote:
> >
> > > echo -e and \e are not POSIX. Depending on what /bin/sh is, we can get
> > > incorrect output like:
> >
> > I'm curious to which shell this is.
>
> Quite frankly I don't know but that's the output we get when we run it in
> Jenkins. I'll try to find out.
The only shell that did not support \e was dash -- I fixed it myself on
2017-01-24; thus I expect any Ubuntus prior to Zesty to require \033. This
means your Jenkins likely runs Xenial.
\e has been supported since ages in at least: bash zsh mksh sash posh ksh
busybox:sh; also in perl python ruby lua php, gcc clang tcc -- MSVC being
the only other exception I know about.
Indeed POSIX doesn't specify \e, but as it pretends there are charsets other
than ASCII and EBCDIC, it can't. There's no escape in its "portable
character set".
Meow!
--
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁
⢿⡄⠘⠷⠚⠋⠀ Have you accepted Khorne as your lord and saviour?
⠈⠳⣄⠀⠀⠀⠀
Powered by blists - more mailing lists