Turn a github.com URL into a git.io URL.

gitio

#!/bin/sh
# SPDX-License-Identifier: MIT

## Copyright (C) 2011 Przemyslaw Pawelczyk <[email protected]>
##
## This script is licensed under the terms of the MIT license.
## https://opensource.org/licenses/MIT
#
# Usage: gitio URL [CODE]
#
# Turns a github.com URL
#  into a git.io URL
#
# Copies the git.io URL to your clipboard if xclip is available.

URL="$1"
CODE="$2"

if ! expr "${URL}" : "\(\(https\?://\)\?\(gist\.\)\?github.com/\)" >/dev/null; then
	echo "* github.com URLs only" >&2
	exit 1
fi

if ! expr "${URL}" : "http" >/dev/null; then
	URL="https://${URL}"
fi

OUT="$(\
	curl -si https://git.io -F "url=${URL}" ${CODE:+-F "code=${CODE}"} | \
	sed '/^Status: /{s///;/^201/d;q};/^Location: /!d;s///'
	)"

if expr "${OUT}" : "[0-9]\+" >/dev/null; then
	echo "${OUT}" >&2
	exit 1
fi

echo "${OUT}"
which xclip >/dev/null && echo "${OUT}" | xclip -selection clipboard

Based on @defunkt's ruby snippet.

I think something is wrong with the regular expression. I can't seem to get any github.com URL to work…

@pandamonia:
Dunno what's wrong, as regexps are definitely correct. I rechecked it just now:

$ wget -q --content-disposition https://gist.github.com/gists/1275673/download
$ tar zxf gist1275673-*.tar.gz
$ cd gist1275673-835ba0b0d2ab3d2ec48504481acdf36c442e168c/
$ chmod +x gitio
$ ./gitio https://github.com/wycats/handlebars.js handlebars
http://git.io/handlebars

I have xclip, so URL went to the clipboard too. Further checking:

$ ./gitio http://pandamonia.us/
* github.com URLs only
$ ./gitio https://github.com/pandamonia/OHAttributedLabel
http://git.io/jx45wg
$ ./gitio github.com/johnezang/JSONKit/issues/47\#issuecomment-2485870
http://git.io/JtY5FA

Everything works for me.

You meant that you always get * github.com URLs only message?
What OS, shell, expr (which is part of coreutils) do you have?

@przemoc I'm running Lion 10.7.2 and bash. expr is installed…

When I run the three tests you did in the last code block, I get the * github.com URLs only every time.

@pandamonia: Cut out line 13 to 20 (inclusive) and then it will work, am I quite sure of :)

@knarf-se / Re: https://gist.github.com/przemoc/1275673/#comment-63754:

Like the following?

#!/bin/sh    

# Usage: gitio URL [CODE]
#
# Turns a github.com URL
#  into a git.io URL
#
# Copies the git.io URL to your clipboard if xclip is available.

URL="$1"
CODE="$2"

OUT="$(\
    curl -si http://git.io -F "url=${URL}" ${CODE:+-F "code=${CODE}"} | \
    sed '/^Status: /{s///;/^201/d;q};/^Location: /!d;s///'
    )"

if expr "${OUT}" : "[0-9]\+" >/dev/null; then
    echo "${OUT}" >&2
    exit 1
fi

echo "${OUT}"
which xclip >/dev/null && echo "${OUT}" | xclip -selection clipboard    

I know I'm super late to this party, but...

Shouldn't you just update your Regex to match the one written by @defunkt in https://gist.github.com/defunkt/1209316#file-gitio-L12?

So this: \(\(https\?://\)\?\(gist\.\)\?github.com/\)
Becomes this: /^(https?:\/\/)?(gist\.)?github.com/ ?

It appears that there have been a number of updates to the original gist since you created this one.

@rex, being super late is fine, but you're also super wrong (sorry, I couldn't resist...).

@defunkt's gist has only 3 revisions and the last one is only a name change. All his revisions are from September 2011. My gist was created later (October 2011) and originally included regexp anchors (^), but it was useless (or even more like wrong as having different behavior in different implementations), because expr matching does anchored pattern matching anyway, so I removed them (November 2011). Regular expressions additionally differ, because ruby's RE flavor, which is Perl-like and closer to Extended Regular Expressions, is different from expr's one, which is Basic Regular Expressions.

And if you're talking about forks, I haven't checked them. I just looked at the forks of my script right now (only 2 introduced some changes) and I don't see any important stuff there. @rduenasf replaced xclip call with pbcopy (which is apparently MacOS thing) and @skopp introduced README.md and renamed the script twice (I simply don't get why people do such superfluous changes at all, but they're surely free to do whatever they want).

I always appreciate suggestions, comments and questions, but I don't like rushed and showy ones, because they usually lack even the most basic research regarding commented matters.

http://git.io/gitio points here

I pushed very small update - https is used now (it should be used for quite some time already).

commit af12b6b48e6220daab23a92d69957407526aa0a7
Author: Przemyslaw Pawelczyk <[email protected]>
Date:   2018-02-06 14:27:24 +0100

    gitio: Add copyright notice and MIT license notice.

commit 6a7f2ebe37ea089adf6762e5591e8e99fe95d4fe
Author: Przemyslaw Pawelczyk <[email protected]>
Date:   2018-02-06 14:29:32 +0100

    Add SPDX License Identifier.

    The Software Package Data Exchange (SPDX) is a good initiative, it has
    matured over time and deserves accelerated adoption in open-source.

    https://spdx.org/learn
    https://spdx.org/using-spdx
    https://spdx.org/license-list