Help a sed noob

Hi,

There is a construct like this in git-parse-remote.sh which makes busybox ash unhappy:

                sed -ne '/^URL: */{
                        s///p
                        q
                }' "$GIT_DIR/remotes/$1"

It complains about "no previous regexp" while gnu sed is ok. Can anyone explain to me what does "s///p" do? GNU Sed info page says nothing about empty regexp. If I replace it with "s/\(.*\)/\1/p" then I get "URL: " along with the remote path.

By the way, can we use another construct instead? It would be less work for me ;-)

"Nguyen Thai Ngoc Duy" <pclouds@gmail.com> writes:

Traditionally empty LHS regexp means "the same as last match". Replace it with "s/^URL: *//p" and you would be Ok.

Great! Thank you both.

On 8/3/07, Junio C Hamano <gitster@pobox.com> wrote:

Nguyen Thai Ngoc Duy schrieb:

You could manually add the previous regular expression, like this:

		sed -ne '/^URL: */{
			s/^URL: *//p
			q
		}' "$GIT_DIR/remotes/$1"

René

also sprach Nguyen Thai Ngoc Duy <pclouds@gmail.com> [2007.08.04.0122 +0200]:

Looks like it prints the first matching line:

  grep '^URL: *' | head -1

On Sat, Aug 04, 2007 at 12:31:26PM +0200, martin f krafft wrote:

No, it prints everything _except_ "URL: *" from the first matching line.

-Peff

also sprach Jeff King <peff@peff.net> [2007.08.04.1350 +0200]:

I am convinced we could submit this to the Obfuscated Code Contest. :)

The behaviour is absolutely unclear from the manpage and defies my logic. Can you elaborate a bit, even though this is off-topic?

martin f krafft wrote:

This is basic sed scripting.

-ne means "don't print, execute the following code".

/.../ should be clear, { starts a "function list" s// is "match last regexp, so s/// is "match last regexp and replace with empty string. flag "p" to function "s" means "print if matched" q is "quit".

that's really straight forward and all documented in my sed(1).

cheers
  simon

also sprach Simon 'corecode' Schubert <corecode@fs.ei.tum.de> [2007.08.04.1426 +0200]:

This fact is not in the manpage, AFAICT, at least not on Debian systems.

martin f krafft <madduck@madduck.net> writes:

IIRC, this is a traditional _ed_ feature. I don't think it has ever been a traditional part of _sed_, but maybe some implementations have adopted this usage.

On Sat, Aug 04, 2007 at 02:18:11PM +0200, martin f krafft wrote:

The original sed code in question was:

                sed -ne '/^URL: */{
                        s///p
                        q
                }' "$GIT_DIR/remotes/$1"

There are a few things to note:
  1. -n means "do not print lines by default"
  2. sed addresses consist of an address (in this case a regex meaning
     "do this for lines that match the regex") and a command
  3. The braces start a set of commands, so that for lines matching the
     address, we do all of the commands.
  4. An empty matching portion for a regex means "use the last regex".

So this script comes down to:
  - don't write any lines except the ones we match
  - find a line that starts with URL:
    - replace the URL: part with nothing
    - print the result
    - quit

It could be more simply written as:

sed -ne 's/^URL: *//pq'

which uses the substitution as an address, but I don't know whether that was allowed in the original sed.

-Peff