This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc 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 2/2] posix: New Linux posix_spawn{p} implementation


On 28-01-2016 15:22, Mike Frysinger wrote:
> On 27 Jan 2016 08:17, Paul Eggert wrote:
>> Adhemerval Zanella wrote:
>>> +#ifdef O_CLOEXEC
>>> +# ifndef __ASSUME_PIPE2
>>> +  if (__have_pipe2 >= 0)
>>> +# endif
>>> +    {
>>> +      r = __pipe2 (pipe_fds, O_CLOEXEC);
>>> +# ifndef __ASSUME_PIPE2
>>> +      if (__have_pipe2 == 0)
>> ...
>>
>> This sort of code is hard to read. Instead, declare substitutes like this after 
>> you do your #includes:
>>
>>    #ifndef __ASSUME_PIPE2
>>    # define __have_pipe2 1
>>    #endif
>>    #ifndef O_CLOEXEC
>>    # define O_CLOEXEC 0
>>    #endif
>>
>> and let the rest of the code just use __have_pipe2 and O_CLOEXEC, without the 
>> forest of #ifdefs.
> 
> where does O_CLOEXEC not exist ?  can't we clean that up ?
> 
> looking at __ASSUME_O_CLOEXEC, that too is set to 1 everywhere
> (linux/nacl/hurd).  rather than add more code using these, can't
> we delete this old code ?  or at the very least, do not try to
> use those symbols in this new code.
> -mike
> 

According to 74385da564a92fc441a7d7f3ed7a4295b74bfdbf we can now assume
O_CLOEXEC as always existent for Linux. I will change that.


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