Fallback with ExpiresDefault does not seem to include images

[creopard]

Fallback with ExpiresDefault does not seem to include images

[LeoColomb]

Thanks for opening this issue @creopard.
That being said, the behavior you are facing is weird.
Indeed, the unit test framework used to validate the configuration explicitly checks Cache-Control value:

https://github.com/h5bp/server-configs-test/blob/a3dee296da841834aa2d130c1f887751553a5e58/lib/basic-file-access.json#L11

I'm not excluding the option that these tests are weak, but I'd appreciate more investigation on your side.
Are you sure LiteSpeed is following ExpiresDefault same as Nginx?

[creopard]

Are you sure LiteSpeed is following ExpiresDefault same as Nginx?
I've looked up the default config, although I can't access the one used by the webhoster. It seems the default comes from the main/global config file: https://openlitespeed.org/kb/openlitespeed-directive-working-in-progress/#expiresDefault

expiresDefault A604800
expiresByType image/*=A604800, text/css=A604800, and so on...

meaning that everything defaults to 1 week.
I was also wondering where that 1 week setting initially came from.

Assumption: If the value of expiresDefault will be overridden by the .htaccess with access plus 1 year, this seeting does not seem to affect the preset done in the above main config file expiresByType image/*=A604800 but only types that have no default setting yet.

Hence, (in my case) I need to explicitly set expiresByType image/* as well in the .htaccess file order to override the default from the main file.

It would be great if anybody else could also look up the expiry dates of images in a real world example.
Or is it just me? In the latter case, we can close this issue of course ;-)

[LeoColomb]

Assumption: If the value of expiresDefault will be overridden by the .htaccess with access plus 1 year, this seeting does not seem to affect the preset done in the above main config file expiresByType image/*=A604800 but only types that have no default setting yet.

That sounds to be the good assumption 😊
But it is for the same assumption that I'd advocate to avoid uncommenting this section.
It's always easier to raise the granularity when needed than "fixing" the upper specific generics.
I hope it makes sense!

[creopard]

Sure, I would just make people (who can't see the webhosters config, if they are on shared environments) aware that the entry
ExpiresDefault "access plus 1 year"
does not seem to be an automatic fallback scenario for every type that have not been mentioned in the .htaccess explicitly.

[LeoColomb]

@creopard Sounds reasonable. Would you like to submit a PR to add this note?

[creopard]

Added pull request: #371

[LeoColomb]

Thanks!