From: benny AT crocodial DOT de (Benjamin Riefenstahl)
Subject: Re: BUG: Mingw32: _O_TRUNC
7 Apr 1998 08:23:12 -0700
Message-ID: <352117C6.41BA4A6B.cygnus.gnu-win32@crocodial.de>
References: <19980326182829 DOT 16068 DOT qmail AT hotmail DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
To: gnu-win32 AT cygnus DOT com

Hi Earnie,


Earnie Boyd wrote:
> The value defined in fcntl.h for _O_TRUNC is 0x0200 which equates to 512
> decimal.  When used in the open routine it gives an error value 22 which
> equates to "Invalid argument" from perror or defined as
> ERROR_BAD_COMMAND in the Windows32/Errors.h header.

Looking at the source for the VC5 RTL this would happen if you have set
neither _O_WRONLY nor _O_RDWR. The source has this also in some
"This-can't-happen" branches.

> I modified the value of _O_TRUNC to 0x0020 which equates to 32 decimal.
> No error is issued from the open routine.

0x0020 is _O_SEQUENTIAL in VC5. I would guess that's just an optimizing
hint.

> I know the the SDK/misc/posix/fcntl.h defines it as 0x00000200 which is
> the same as 0x0200.  So, does the open routine in crtdll have a bug, or
> is the _O_TRUNC value actually wrong?

What does the NT Posix subsystem have to do with anything here?


so long, benny
======================================
Benjamin Riefenstahl (benny AT crocodial DOT de)
Crocodial Communications EntwicklungsGmbH
Ruhrstraße 61, D-22761 Hamburg, Germany
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".