Re: [Mingw-users] Problem with GCC not outputting cr-lf for Windows

[Paul <pau...@af...>]

Sorry, my previous mail got sent by accident before it was completed
(in case it eventually shows up on this list, it hasn't as of
completing this mail).

As I was saying...

On Tue, 28 Aug 2012 23:34:35 +0100
Keith Marshall <kei...@us...> wrote:

> On 28/08/12 22:48, Stuart Soloway wrote:
> > It turns out that I was invoking the executable from the MinGW
> > shell.
> 
> I think you are confused.  There is no such thing as "MinGW shell";
> I assume you mean the MSYS shell.
If I look in the Start menu on my computer, under MinGW is "MinGW
Shell", which of course launches the MSYS shell. Can't blame people for
calling it what's written on the box. Dunno, maybe current
versions have changed this, but I'm guessing enough people still see
this.

Incidentally, this may also be part of the reason for the continuous
confusion over MSYS vs. MinGW that people seem to have.

> > When invoked that way, it uses Linux line-termination.  But if I
> > invoke it from the Windows command line it uses Windows line
> > termination.
> 
> Is this a native Windows build, compiled with the MinGW compiler?
> If so, then it does no such thing; it writes \r\n regardless of
> which shell you launch it from.
> 
> See, here's a thing about text mode files, written and subsequently
> read back using *Microsoft's* MSVCRT.DLL I/O, (which is what MinGW
> uses): when writing, \n is converted to \r\n; when reading, \r\n is
> converted back to \n.
I'm not sure I get this, he said he tested with emacs, so surely emacs
is reporting what is actually in the file, just like od does? Or does
emacs also read it in in text mode, using MSVCRT.DLL? Then surely
notepad does as well? If so, how does MSVCRT.DLL know when to convert
\r\n to \n, and when not to? Because it sure doesn't do it for all
files. I know I've seen many files in notepad that have proper \r\n
line endings. Also, why would it seem to be different based on which
shell he launches the test program from?

> It isn't.  Not in the way you seem to think.  You are confused.
Actually, he seems to be asking the right questions, and the
conclusions he is drawing from his investigations seem perfectly valid
with the results he is observing.

Paul