Post GIST

pgist.sh

# 0. Your file name
FNAME=some.file

# 1. Somehow sanitize the file content
#    Remove \r (from Windows end-of-lines),
#    Replace tabs by \t
#    Replace " by \"
#    Replace EOL by \n
CONTENT=$(sed -e 's/\r//' -e's/\t/\\t/g' -e 's/"/\\"/g' "${FNAME}" | awk '{ printf($0 "\\n") }')

# 2. Build the JSON request
read -r -d '' DESC <<EOF
{
  "description": "some description",
  "public": true,
  "files": {
    "${FNAME}": {
      "content": "${CONTENT}"
    }
  }
}
EOF

# 3. Use curl to send a POST request

# ANONYMOUS GIST :
# curl -X POST -d "${DESC}" "https://api.github.com/gists"

# REGISTERED USER
curl -u "${GITHUB_USERNAME}" -X POST -d "${DESC}" "https://api.github.com/gists"

If you need to create an anonymous gist comment the last line and uncomment:

curl -X POST -d "${DESC}" "https://api.github.com/gists"

posted on http://stackoverflow.com/a/33354920/1069083

I don't have enough stackoverflow reputation to add a comment to the main conversation thread, but the reason why this breaks if the file contains a percent symbol (%) is because awk then attempts to treat it as a "Format-Control Letter" - please see https://www.gnu.org/software/gawk/manual/html_node/Control-Letters.html HTH, Jaime

And it turns out that the solution is just to "double-escape" the percent symbols in the input, so line 9:
CONTENT=$(sed -e 's/\r//' -e's/\t/\t/g' -e 's/"/\"/g' "${FNAME}" | awk '{ printf($0 "\n") }')
becomes:
CONTENT=$(sed -e 's/\r//' -e's/\t/\t/g' -e 's/"/\"/g' "${FNAME}" | awk '{gsub("%","%%",$0);print $0}' | awk '{ printf($0 "\n") }')
:-)

This code deserves a repo: https://github.com/ceremcem/create-gist (I'll hand it over to @s-leroux on request)