Re: Problems with echo. (echo period)

[Eli Zaretskii]

> From: "Dave Korn" <address@hidden>
> Cc: <address@hidden>,
>       <address@hidden>
> Date: Tue, 24 Apr 2007 00:22:32 +0100
> 
> On 23 April 2007 23:02, Eli Zaretskii wrote:
> 
> >> From: "Dave Korn" <address@hidden>
> >> Cc: <address@hidden>,
> >>    <address@hidden>
> >> Date: Mon, 23 Apr 2007 15:22:40 +0100
> >> 
> >>> I don't think "echo." is a shell builtin.  It is a peculiar feature of
> >>> the cmd.exe command parser.
> >> 
> >>   This is a semantic quibble.
> > 
> > You are, of course, entitled to think so.  But your thinking so
> > doesn't yet make it so.
> 
>   No, it doesn't; the fact that no external executable is invoked and the
> entire functionality is supplied by a subroutine within cmd.exe, on the other
> hand, /does/ make it so

We are talking past each other.  I reacted to the suggested change to
add "echo." to the table of built-ins.  It is in that context that you
should read my ``I don't think "echo." is a shell builtin'': I meant
to say that, since "echo." is not a built-in, it is wrong to solve
this in a kludgy way by adding "echo." to the table of built-ins.

> the point remains that make attempts
> to invoke the slow path for shell builtins, and what the command line
> specifies should occur here is the invocation of a shell builtin, and make
> fails to identify it as such a case and fails to take the slow path and fails
> to cause the shell builtin, however you may name it, to be invoked.

I agree with this analysis.  In short, Make fails to follow the
command-line parsing rules of cmd, and thus behaves differently from
what one would see at the cmd prompt.

>   In any case, we are in agreement that it's not worth attempting to emulate
> the general weirdness of cmd.exe parsing, so the argument is entirely
> academic.  Your turn!

Being in agreement doesn't have to rob us of the delight of continuing
the argument!