This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [PATCH 3/3] libctf: look for a qsort_r that functions properly

From: Nick Alcock <nick dot alcock at oracle dot com>
To: Pedro Alves <palves at redhat dot com>
Cc: binutils at sourceware dot org, jose dot marchesi at oracle dot com, John dot W dot Marshall at glasgow dot ac dot uk, sebastian dot huber at embedded-brains dot de
Date: Wed, 05 Jun 2019 11:43:03 +0100
Subject: Re: [PATCH 3/3] libctf: look for a qsort_r that functions properly
References: <20190603204735.203510-1-nick.alcock@oracle.com> <20190603204735.203510-3-nick.alcock@oracle.com> <0e508b57-7687-f71c-cfd5-306884d9f694@redhat.com>

On 5 Jun 2019, Pedro Alves told this:

> On 6/3/19 9:47 PM, Nick Alcock wrote:
>
>> We cannot just look for any declaration of qsort_r, because some
>> operating systems
>
> OOC, could you give some examples?

This is a BSD-heritage thing, so MacOS X and FreeBSD are examples.

>> So, instead, use AC_RUN_IFELSE to check for the function of qsort_r(),
>> passing function pointers in to both args and making sure that we never
>> actually dereference the arg pointer.
>
> AC_RUN_IFELSE is really best avoided; obviously you can't run when
> cross compiling.
>
> How does gnulib handle this?

This is an old patch: v2, posted later, uses gnulib's method (and mostly
gnulib's code too).

-- 
NULL && (void)

References:
- [PATCH 1/3] libctf: fix the type of ctf_enum.cte_value
  - From: Nick Alcock
- [PATCH 3/3] libctf: look for a qsort_r that functions properly
  - From: Nick Alcock
- Re: [PATCH 3/3] libctf: look for a qsort_r that functions properly
  - From: Pedro Alves

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]