[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: automake-1.11.1 test failure on Mac OS X 10.6
From: |
Rugxulo |
Subject: |
Re: automake-1.11.1 test failure on Mac OS X 10.6 |
Date: |
Sat, 16 Jan 2010 06:46:22 -0600 |
Hi,
On Sat, Jan 16, 2010 at 5:46 AM, Ralf Wildenhues <address@hidden> wrote:
> Hello Rugxulo, Andris,
>
> I need your help for a bit of Autoconf testing for me; see below.
>
> This is where said patch went wrong: the logic never meant to use '\r'
> which some sed's but not all expand to a literal carriage return; IOW,
> we were relying on, and fooled by, undefined behavior in sed here.
>
> Can you please try the following: grab Autoconf 2.65 from ftp.gnu.org,
> apply the patch below, build and install it on DJGPP, and see whether
> it works on this example package?
I'm not sure 2.65 will even build with DJGPP, that's the problem. At
least, not without some work, and it's probably beyond me. I can maybe
try building later, though.
> Fix substitution of carriage return on Darwin.
> - ac_cs_awk_cr='\r'
> + ac_cs_awk_cr='\\r'
Here's what I want to know:
1). What sed (vendor, version) does Mac OS X/Darwin use? *BSD or just
really old GNU?
2). Is it "\r" or "\\r" that is really undefined? I was never sure if
"\\" was supposed to be undefined in POSIX (e.g. gnused --posix) or
not.
Hmmm, latest sed.info (4.2.1) says this:
Note that the only C-like backslash sequences that you can
portably assume to be interpreted are `\n' and `\\'
Somebody should probably ask Paolo Bonzini. I'll cc him. :-)
- Re: automake-1.11.1 test failure on Mac OS X 10.6,
Rugxulo <=