The .replit file gets autofilled with a ton of garbage when Whoogle is
imported, including a required "entrypoint" field that defaults to
"main.py" (even though the run and onBoot fields were already included
and should negate the need to specify an entrypoint, but whatever).
I'm not going to restructure Whoogle to fit what Replit wants, so I've
moved the startup commands to their own script (misc/replit.py) and
updated the "entrypoint" field in .replit.
Setting `WHOOGLE_TOR_FF` will attempt to modify the existing torrc file
with the "FascistFirewall 1" config, which restricts outbound ports to
reachable values.
Fixes#907
Still a manual process, but better than updating the json line by line.
I'm undecided on automating this anytime the translation json gets
updated, since it could possibly ratelimit lingva instances.
The Tor service logs often confuse Whoogle users, since they're a lot
more verbose than anything Whoogle ever reports. The bulk of these logs
use "notice" level logging and are not helpful for the average user, so
everything between debug and notice is now directed to /dev/null.
Fixes#825
Allows skipping the Tor startup script if WHOOGLE_TOR_SERVICE is set to
0. This is separate from WHOOGLE_CONFIG_TOR, which only allows
enabling/disabling user configuration of passing searches through
Tor.
Closes#631
Added password authentication for tor control port.
For user configuration of access to tor control port. This file should be
heavily restricted in file system.
Co-authored-by: MadcowOG <madcowog@Arch-Main.localdomain>
Depending on bash wasn't strictly necessary, as the two minimal scripts
in the repo were both nearly POSIX anyways.
Aside from simplifying the repo's dependencies a little bit, this also
helps reduce the overall Docker image size as an added bonus.
Including a list of instances that are easily machine-readable allows
services such as Farside (https://github.com/benbusby/farside) to read
these and have an up to date list of valid instances.
Introduces a new script for quickly regenerating a Heroku instance
(typically with a new IP) to handle the rare circumstances where an
instance is flagged by Google and prompted for a captcha.
The config/ dir was renamed to misc/ to avoid confusion with the Docker
mounted config volume, and to more closely match its intended purpose
(which is to contain all miscellaneous features/scripts/etc that add
functionality to Whoogle, but are not critical for general use).
Pip installs of whoogle search were missing access to the misc/ folder,
which previously contained the language and country json files. These
have been moved to app/misc, and the previous root level misc/ was
renamed to config/ (since it now only contains the tor config files).
Bump to 0.3.1.
Moves the language and country dicts from the config model to json files
that are loaded during app init and stored in the app config dict. This
substantially improves the readability of the config model and allows
for much more sensible loading of the language/country options.
The tor service is now started by calling a script which runs tor
according to the current container user. If the user is root, the script
will begin the tor service as normal. Otherwise, it runs tor as the
current user. This primarily is meant to address the issue with Heroku
builds (which don't have a root user) not being able to start tor as a
service.
Also refactored the rc/ dir to misc/ (with a tor/ subdir) since that
makes more sense.
