| Commit message (Collapse) | Author | Age | Files | Lines |
| | |
|
| | |
|
| |\
| |
| |
| |
| |
| | |
store + Prevent voting when not allowed' (#552) from davidrv00/snac2-fork:vote-fix into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/552
|
| | | |
|
| | | |
|
| | | |
|
| |\ \
| | |
| | |
| | |
| | |
| | | |
davidrv00/snac2-fork:poll-mastoapi into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/551
|
| | |/ |
|
| |\ \
| | |
| | |
| | |
| | |
| | | |
(#549) from oxzi/snac2:strip-metadata-man-warning into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/549
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add words of caution and reasoning to the "strip_exif" configuration for
the server.json file, as these commands would be executed outside of the
sandbox - at least on OpenBSD - and both have quite a history on
security issues due to their huge attack surface and variety of
supported protocols.
After getting comfortable with the related code, I would continue using
it on a personal instance, but would reconsider enabling "strip_exif" on
shared instances with multiple users.
IMO, snac administrators should at least know of potential dangers.
|
| | | | |
|
| |\ \ \ |
|
| | |\ \ \
| | |_|/
| |/| |
| | | |
| | | |
| | | | |
oxzi/snac2:strip-metadata-openbsd into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/548
|
| | | |/
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Change the strip_exif logic to work with the already existing OpenBSD
sandbox and allow ffmpeg and mogrify to be executed.
The previous strip_exif implementation relied on system(3), effectively
starting "/bin/sh" and executing the required tool within a shell
session. Making this work in the sandbox would require to allow
executing "/bin/sh", rendering the sandbox useless.
Thus, the code now starts determining the absolute path of the tools -
unless they are given as ffmpeg_path or mogrify_path - and allowing them
to be executed via unveil(2). Then, instead of the system(3) call, the
good old fork(2) and execve(2) dance is performed.
The sbox_enter code was made aware of strip_exif, which resulted in a
pledge(2) violation before when disable_email_notifications was set to
false. Furthermore, the detected paths of the tools are now allowed.
|
| |/ / |
|
| |/ |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |\
| |
| |
| | |
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/546
|
| | | |
|
| | | |
|
| |\|
| |
| |
| | |
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/545
|
| |/ |
|
| |\ |
|
| | | |
|
| | | |
|
| |/
|
|
|
|
| |
Instead of unconditionally stripping ? variables when retrieving the keyId,
try first calling actor_request() directly, and only strip them and retry
if it fails.
|
| | |
|
| | |
|
| |\
| |
| |
| |
| |
| | |
daltux/snac2:pt_BR-8 into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/543
|
| |/ |
|
| | |
|
| | |
|
| |\
| |
| |
| |
| |
| | |
zen/snac2:master into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/539
|
| | | |
|
| |/ |
|
| |\
| |
| |
| |
| |
| | |
byte/snac2:notif_filter into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/524
|
| | | |
|
| | | |
|
| |\ \
| | |
| | |
| | |
| | |
| | | |
into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/538
|
| | |/ |
|
| |\ \
| |/
|/|
| |
| |
| | |
daltux/snac2:pt_BR-7 into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/537
|
| |/ |
|
| | |
|
| |\
| |
| |
| |
| |
| | |
(#530) from dandelions/snac2:exact-post-times into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/530
|
| | | |
|
| | | |
|
| |\ \
| | |
| | |
| | |
| | |
| | | |
dandelions/snac2:poll-limits into master
Reviewed-on: https://codeberg.org/grunfink/snac2/pulls/534
|