Re: [PATCH] On non-Cygwin, use O_TEXT when O_BINARY is not set.

Michael Haubenwallner <[email protected]> Thu, 21 Feb 2019 13:12:46 +0100

Newsgroups	gmane.comp.lib.newlib
Message-ID	<[email protected]>


On 2/21/19 10:56 AM, Corinna Vinschen wrote:
> On Feb 21 09:08, Michael Haubenwallner wrote:
>> * libc/stdio/flags.c (__sflags): Fix typo when testing if O_BINARY is
>> set, to really set O_TEXT if not O_BINARY, applies to non-Cygwin only.
>> ---
>>  newlib/libc/stdio/flags.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/newlib/libc/stdio/flags.c b/newlib/libc/stdio/flags.c
>> index 71fc1f60a..cd73385ce 100644
>> --- a/newlib/libc/stdio/flags.c
>> +++ b/newlib/libc/stdio/flags.c
>> @@ -90,7 +90,7 @@ __sflags (struct _reent *ptr,
>>  	}
>>      }
>>  #if defined (O_TEXT) && !defined (__CYGWIN__)
>> -  if (!(m | O_BINARY))
>> +  if (!(m & O_BINARY))
>>      m |= O_TEXT;
>>  #endif
>>    *optr = m | o;
>> -- 
>> 2.19.2
> 
> Given that O_TEXT is under #if defined (__CYGWIN__) guard, do
> we need this code at all?  Does anybody actually *want* O_TEXT
> outside of the Windows world?

Actually I've no idea whereever newlib can go beyond Cygwin and Linux.
To me, this line is a typo, and I won't care if these lines are dropped.

Thanks!
/haubi/

signature.asc (application/pgp-signature, 981 B)