Add Content-Security-Policy header #2123

Closed
tae wants to merge 0 commits from gitlab-mr-iid-856 into develop
Member

Now dev testing can reproduce the security issues we saw in production.

Now dev testing can reproduce the security issues we saw in production.
Member

looks ok to me

looks ok to me
Owner

i don't really understand why tho

i don't really understand why tho
Owner

It's really painful when devs spend time working on things and then they submit an MR and everything looks OK but you get security issues breaking the change/feature in production because the CSP headers are denying things that nobody saw in dev. The sooner the devs can catch these errors the better, I think.

It's really painful when devs spend time working on things and then they submit an MR and everything looks OK but you get security issues breaking the change/feature in production because the CSP headers are denying things that nobody saw in dev. The sooner the devs can catch these errors the better, I think.
Owner

I just had a new feature testing in mfc-develop which was great until i deployed it to dev server and got a white screen because CSP violation that developer did not see because it was never deployed to a production Pleroma instance with the CSP headers enabled...

edit: merged this into mfc-develop branch for our use, but I encourage us to do this upstream as well so these issue can be caught as early as possible in the development process

I just had a new feature testing in `mfc-develop` which was great until i deployed it to dev server and got a white screen because CSP violation that developer did not see because it was never deployed to a production Pleroma instance with the CSP headers enabled... edit: merged this into `mfc-develop` branch for our use, but I encourage us to do this upstream as well so these issue can be caught as early as possible in the development process
Owner

I get the feeling your broken feature used something weird like iframe from another domain which we most likely wouldn't accept upstream...

Still, real solution would be (optional) copying CSP headers from BE since those could change on BE side or instance in question might have its own CSP headers defined in nginx - the old way.

I get the feeling your broken feature used something weird like iframe from another domain which we most likely wouldn't accept upstream... Still, real solution would be (optional) copying CSP headers from BE since those could change on BE side or instance in question might have its own CSP headers defined in nginx - the old way.
Owner

I see no harm in having strict ones in the FE by default

I see no harm in having strict ones in the FE by default
Owner

i mean sure, i just wanted to know why

i mean sure, i just wanted to know why

Pull request closed

Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
5 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
pleroma/pleroma-fe!2123
No description provided.