Setting the MIX_ENV variable in rc_pre() isn't possible, because the
environment doesn't persist between rc_pre and rc_start(). This way we
can also ditch the custom rc_start() function in favor of the default
one which is just:
rc_start() {
rc_exec "${daemon} ${daemon_flags}
}
OpenBSD's httpd does not support caching in any way and putting a
caching layer between it and Pleroma is pointless when nginx works fine.
I also ran into issues with relayd when accessing it from the Tor browser.
Federation seems to be unaffected by this as is base Firefox and Chrome.
* I cretaed a folder 'development'
* I split up the file dev.md into three parts and moved it to this folder
* index.md
* authentication_authorization.md
* mrf.md
* I also moved ap_extensions.md
* I created a new file setting_up_pleroma_dev.md
Added ffmpeg/imagemagick checks to launch checks (if media preview proxy is enabled). Added documentation on installing optional media / graphics packages (imagemagick, ffmpeg, exiftool).
* I added an include and use this include for the installation guides that already had this section
* I added the "Further reading" section as well as te "Questions" section to the English guides that didn't have it yet
* I added a first point "How Federation Works/Why is my Federated Timeline empty?" to link to lains blogpost about this because we still get this question a lot in the #pleroma support channel
* I reordered the list a bit
* initdb command now requires explicit -U and -D flags
* acme-client now requires version v02 API endpoint
* Added step to add admin user, per other installation guides
* Corrected setup_db.psql path
I understand that this change is quite unreadable and hard to review, sorry for forgetting to do atomic commits. This patch does not change too much content wise, it just
* Gets everything from gitlab wiki
* Removes some specific gitlab hacks
* Formats all documentation file names to be in snake case so they look the same way as our code does
