Use string() in XPath to avoid further sed-e-s and just extract the 1st og:imag… - plumb - Open certain URL patterns with an ad-hoc opener (plumber)
 (HTM) hg clone https://bitbucket.org/iamleot/plumb
 (DIR) Log
 (DIR) Files
 (DIR) Refs
 (DIR) README
       ---
 (DIR) changeset e89f98c6d54347754d8d8abbf4084850717658d7
 (DIR) parent d86a99e60c70cae3df952dbacb69305d60ad77e5
 (HTM) Author: Leonardo Taccari <iamleot@gmail.com>
       Date:   Tue, 14 Aug 2018 02:15:13 
       
       Use string() in XPath to avoid further sed-e-s and just extract the 1st og:image.
       
       Diffstat:
        openers/ogimage |  4 ++--
        1 files changed, 2 insertions(+), 2 deletions(-)
       ---
       diff -r d86a99e60c70 -r e89f98c6d543 openers/ogimage
       --- a/openers/ogimage   Mon Aug 06 23:44:35 2018 +0200
       +++ b/openers/ogimage   Tue Aug 14 02:15:13 2018 +0200
       @@ -2,7 +2,7 @@
        
        for u in "$@"; do
               imgurl=$(curl -s -L "$u" |
       -           xmllint --html --xpath '//meta[@property="og:image"]/@content' - 2>/dev/null |
       -           sed -e 's/^ content="//' -e 's/"$//' -e 's/\?.*$//' -e 's/:large$//' )
       +           xmllint --html --xpath 'string(//meta[@property="og:image"][1]/@content)' - 2>/dev/null |
       +           sed -e 's/\?.*$//' -e 's/:large$//' )
               plumb "${imgurl}"
        done