[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: bug#17773: AIX build errors with coreutils-8.22

From: Pádraig Brady
Subject: Re: bug#17773: AIX build errors with coreutils-8.22
Date: Sat, 14 Jun 2014 19:10:25 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130110 Thunderbird/17.0.2

On 06/14/2014 06:08 PM, Paul Eggert wrote:
> I looked into this some more, and found some problems:
> * On Solaris we also need to define _POSIX_PTHREAD_SEMANTICS to prefer POSIX 
> to pre-POSIX semantics for a few functions; see 
> <http://docs.oracle.com/cd/E19683-01/806-6867/compile-fig-5/index.html>.  The 
> extensions module does this, so a simple fix is to depend on the extensions 
> module.  Almost everybody uses extensions anyway so I don't view this as 
> being a big issue.  (Maybe threadlib should also depend on extensions, 
> instead its calls to AC_USE_SYSTEM_EXTENSIONS and AC_GNU_SOURCE?)
> * The patch uses $host_os so it should AC_REQUIRE([AC_CANONICAL_HOST]).  But 
> why bother?  AC_CANONICAL_HOST is expensive and not every project needs it.  
> Instead, just define the symbols on all platforms; that's what 'extensions' 
> does.
> * This new behavior should be done in gl_EARLY, as _REENTRANT and 
> _THREAD_SAFE affect the visibility of many other symbols.
> So, how about the attached patch instead?  (I haven't tested it.)

Looks good thanks,
though it would be worth mentioning in the commit why extensions is used.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]