curl to get the JSON response for the latest releasegrep to find the line containing file URLcut and tr to extract the URLwget to download itcurl -s https://api.github.com/repos/jgm/pandoc/releases/latest \
| grep "browser_download_url.*deb" \
| cut -d : -f 2,3 \
| tr -d \" \
| wget -qi -
Adding my variant for downloading and installing latest .tar.gz on Linux for oddly named package names:
curl -fsSL "$(curl -fsSL https://api.github.com/repos/aquasecurity/trivy/releases/latest | jq -r '.assets[] | select(.name? | match("tar.gz$")) | .browser_download_url' | grep Linux-$([[ $(uname -m) = "aarch64" ]] && echo "ARM64" || echo "64bit"))" | tar -xz -C /usr/local/bin --A better solution can only come if GitHub allows to specify variants like OS, architecture etc. as part of the release asset upload process.
That would be another topic, on another thread: "how do I download an artifact if the filename doesn't change."
The case we are discussing is very clear and it's pretty different.
And it's really unimportant if you use
sed,grep,jq, because everything woks. Maybe one pipe should be enough, but if you are not seeking for extreme optimization, it's purely stylistic.For instance I rather like the AWK approach:
again, we are hundreds of people, not having much else todo, and playing to find different solutions, but we're not hundreds of idiots. This thread is 8 year old and if there was a simpler solution someone else would have found it in 8 years.