More crash Re: More incmdpos stuff Re: Crash Re: Trial for 5.0.7

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

On Oct 3,  1:27pm, Peter Stephenson wrote:
} Subject: Re: More incmdpos stuff Re: Crash Re: Trial for 5.0.7
}
} On Fri, 03 Oct 2014 13:47:25 +0200
} Mikael Magnusson <mikachu@xxxxxxxxx> wrote:
} > % () ( echo hi; echo $@ ) ls
} > hi
} > ls --color=auto -T 0 -A -v --quoting-style=shell
} 
} I think the right answer is along the lines of "WHY THE **@*@*@ WOULD
} ANYBODY USE SYNTAX LIKE THAT SHELL SYNTAX IS UNREADABLE ENOUGH AS IT IS
} WITHOUT (FROTH, GIBBER, ...)", but instead here's a minimal change until
} I get around to investigating why I thought I needed to leave incmdpos
} = 1 for non-anymous functions.

Well, here's a clue:  With the patch from 33343 applied, it now makes no
difference whether I do or do not back out the parse.c hunks from 32552.
Either way, 'print foo | () cat' crashes.