xfs
[Top] [All Lists]

Re: [PATCH xfsprogs 01/14] configure: use AC_SYS_LARGEFILE

To: Felix Janda <felix.janda@xxxxxxxxx>
Subject: Re: [PATCH xfsprogs 01/14] configure: use AC_SYS_LARGEFILE
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Fri, 12 Aug 2016 13:23:52 -0700
Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20160812165440.GA367@nyan>
References: <cover.1470555003.git.felix.janda@xxxxxxxxx> <ea8cca9bf1dbeb13a788f85f73d6bafbcc374a89.1470555003.git.felix.janda@xxxxxxxxx> <20160809073611.GC25647@xxxxxxxxxxxxx> <20160809174105.GA5907@nyan> <20160812025753.GD10701@xxxxxxxxxxxxx> <20160812165440.GA367@nyan>
User-agent: Mutt/1.6.1 (2016-04-27)
On Fri, Aug 12, 2016 at 06:54:40PM +0200, Felix Janda wrote:
> glibc: since version 2.2 (2000)
> uClibc: since version 0.9.11 (2002)
> dietlibc: since version 0.8 (2001)
> klibc: AFAICS since beginning only transparent LFS
> musl: since beginning only transparent LFS
> bionic: In 2015 _FILE_OFFSET_BITS was implemented "mostly"...
> newlib: (except on cygwin) does not seem to have support for transparent LFS
> 
> Note that LFS can be configured out of uClibc. However its headers
> error out when it is configured out and an application sets
> _FILE_OFFSET_BITS. (So in the case of xfsprogs it would have errored
> out in this situation already earlier.)
> 
> 
> So it seems that this patch series breaks newlib support...

I think that's fine, newlib hasn't every really been a supported
config.  The point I tried to make is that we should aim to error
out during ./configure for this case.

<Prev in Thread] Current Thread [Next in Thread>