Don't touch HTTP headers where possible

[nate-at-gusto]

Even if disable_caching is on, we still overwrite the applications HTTP cache headers.

This can make HTTP caching difficult to observe in development mode. I understand why we have hacks and workarounds here, but it would be better if we could avoid touching this stuff and let the app do it's thing.

[suryanarayanan035]

Hello @nate-at-gusto , we are modifying the HTTP header to prevent browser from storing cache only if disable_caching is on. So, I don't find any problems with adding no-store to HTTP header if disable caching is on and feel like its the right thing to do.

If you still find that we need to prevent modifying HTTP headers, then can you please explain me the problem we will be facing here if we just let it be?

[glaszig]

adding no-store to HTTP header if disable caching is on

disable_caching controls whether no-store is added, yes, but still the header will be overridden, regardless, with must-revalidate, private, max-age=0:

rack-mini-profiler/lib/mini_profiler.rb

Lines 432 to 435 in 5e42a57

	headers['X-MiniProfiler-Original-Cache-Control'] = headers['Cache-Control'] unless headers['Cache-Control'].nil?
	headers['Cache-Control'] = "#{"no-store, " if config.disable_caching}must-revalidate, private, max-age=0"

i think there should also be a big fat warning at boot time when disable_caching == true. i just spent an hour or two digging around rails and rack to see wth i'm doing wrong.

please explain me the problem we will be facing here if we just let it be

the headers which I MYSELF have set are silently overridden and have no effect.

[Michoels]

I just got badly burned by this, for over an hour 🤕

I eventually suspected rack-mini-profiler was the culprit when I noticed the mysterious presence of the x-miniprofiler-original-cache-control header.

@glaszig's suggestion of a "big fat warning" at boot time would be wonderful.

(thanks for the amazing lib, btw!)