How to permalink to a gist's raw file

README.md

Problem: When linking to the raw version of a gist, the link changes with each revision.

Solution:

To return the first file from a gist: https://gist.github.com/[gist_user]/[gist_id]/raw/

• Example: https://gist.github.com/atenni/5604522/raw/
• Works even when you change the filename.

To get a file from multi–file gist: https://gist.github.com/[gist_user]/[gist_id]/raw/[file_name]

• Example: https://gist.github.com/atenni/5604522/raw/README.md

(via: http://stackoverflow.com/a/14529686)

You can use query parameters to get around caching if you want fast updates, i.e. every time you make a request append something like ?cache-bust=(some random url-safe string) to the URL. As far as I can tell, any query string works, so you can omit the cache-bust= if you want, or use donkey-poo= if you feel so inclined.

@rjb3977, usually, an underscore and the current timestamp are used for this: ?_=1617443829.

I found this discussion after looking into it myself. Here's what I've got:

• GIST is NOT meant to behave like a repository. It is a Front End technology.
• Without cache busting, updates take quite a while... Longer than I was going to wait to measure how long exactly.
• WITH cache busting, you can reduce the latency of updates to ~30-60s, but there IS STILL LATENCY.

Here's what I used to observe and measure changes.
Note that I was adding a, b, c... to line 4 character 1 (after the pound/hash sign) every time I saw the change in the cli.

while true; do line=$(curl -s https://gist.githubusercontent.com/raw/2bf5c315a7895193ddc3fe647ef3976b/detect-ci.sh?cache-bypass=$(date +%s) | sed '4q;d'); echo "$(date) -- ${line:1:1}"; sleep 1; done

So, this should be good enough for MOST applications. Mileage may vary.
And there's always a risk that GH will change their caching behavior for these assets.

Nice. Exactly what I needed, thank you! 🙏
I think this should be the default URL, or GitHub should at least be more explicit about this difference...