Re: [PATCH 1/5] chainlint: match arbitrary here-docs tags rather than hard-coded names

On Thu, Aug 09, 2018 at 01:58:05AM -0400, Eric Sunshine wrote:

On Wed, Aug 8, 2018 at 6:50 PM Jeff King [off-list ref] wrote:

quoted

On Tue, Aug 07, 2018 at 04:21:31AM -0400, Eric Sunshine wrote:

quoted

+# Swallowing here-docs with arbitrary tags requires a bit of finesse. When a
+# line such as "cat <<EOF >out" is seen, the here-doc tag is moved to the front
+# of the line enclosed in angle brackets as a sentinel, giving "<EOF>cat >out".

Gross, but OK, as long as we would not get confused by a line that
actually started with <EOF> at the start.

It can't get confused by such a line. There here-doc swallower
prepends that when it starts the swallowing process and removes it add
the end. Even if a line actually started with that, it would become
"<EOF><EOF>cmd" while swallowing the here-doc, and be restored to
"<EOF>cmd" at the end. Stripping the "<EOF>" is done non-greedily, so
it wouldn't remove both of them. Likewise, non-greedy matching is used
for pulling the "EOF" out of the "<...>" when trying to match against
the terminating "EOF" line, so there can be no confusion.

Thanks. I figured you probably had thought of that, but it seemed easier
to ask than to wade through the sed code (I do feel like a bad person to
give that answer, because IMHO one of the key things that makes open
source work is a willingness to dig in yourself rather than asking; but
I am making an exception for this sed code).

Yeah, I was going with the tighter uppercase-only which Jonathan
suggested[1], but I guess it wouldn't hurt to re-roll to allow
lowercase too.

[...]

No. I've gotten so used to \EOF in this codebase that it didn't occur
to me to even think about 'EOF', but a re-roll could add that, as
well.

Thanks. I could take or leave such fixes, since I think our style
discourages both, so I'll leave it up to you whether you want to pursue
them.

-Peff