On 01/02/2019 09:56, Robert Elz wrote:
Date: Fri, 1 Feb 2019 08:00:16 +0000
From: Harald van Dijk <a...@gigawatt.nl>
Message-ID: <5cd22a9c-b213-8ffb-da45-f71057eba...@gigawatt.nl>
| I was referring to the second token on the second line, which is foo,
| not foo=bar.
Oh, sorry, I was in too much of a rush this morning, and did not
read all that carefully (or really, at all...)
For that I might, if some different wording to what is currently
proposed is needed, use something like:
* the TOKEN is the first token observed, or follows other
tokens which would allow this TOKEN to become the command
word of a simple command [xref], should the (unrecognised
yet) following tokens (if any) allow the token sequence
to be parsed that way, but regardless of what tokens follow.
That does not look right: this bullet point is the one that requires
shells to not perform alias substitution on reserved words (since they
would not be parsed as the command name word of a simple command). Your
proposed wording makes this conditional on whether or not the token is
the first in the input stream.
Cheers,
Harald van Dijk