This is a refactor of the codebase to use a more memory-efficient
representation of notes. It should also be much faster at decoding since
we're using a custom C json parser now.
Changelog-Changed: Improved memory usage and performance when processing events
This is a non-behavioral change in preparation for the actual switchover
from Strings to Ids. The purpose of this kit is to reduce the size of
the switchover commit which is going to be very large.
Since we can't mutate NdbNotes, let's update the existing codebase to
generate and sign ids on NostrEvent constructions. This will allow us to
match NdbNote's constructor
Since we don't show these on events anymore, we don't need to spam nip05
validation. We can just check when we go to the profile page
Changelog-Fixed: Eliminate nostr address validation bandwidth on startup
Add the ability to resubscribe to home filters so that it will be
updated when you follow and unfollow people
Changelog-Fixed: Hide users and hashtags from home timeline when you unfollow
I noticed we are not using the PostBox when following new users. Not
good! This is probably why following users sometimes does not work.
Changelog-Fixed: Fixed a bug where following a user might not work due to poor connectivity
Switch the post parser to use the same code as the content parser. This
was causing many issues, including performance issues.
Changelog-Fixed: Fix lag when creating large posts
Changelog-Fixed: Fix npub mentions failing to parse in some cases
Changelog-Added: Add r tag when mentioning a url
Changelog-Removed: Remove old @ and & hex key mentions
This is needed for longform events. Right now we treat unseparated note
artifacts as a list of blocks, but we will likely need to render these
blocks into lists of attributed texts with image blocks inbetween.
The fake note zap attack made me realize that there is a way to do fake
profile zaps using a similar technique. Since damus only checks the
first ptag if it is a profile zap, this means you could include multiple
ptags, the first one being the fake profile with the fake zapper, and
the second p tag as the real target.
This would allow a fake zapper to create a fake a zap, while the zap
notification would still appear for the second ptag because damus
listens for zap events via #p, and that would match the second ptag.
To fix this, ensure that zaps only have at most 1 ptag and 0 or 1 etag.
my CLN zapper checks this but if we don't check this here as well then
we run into fake zap issues.
Changelog-Fixed: Fix potential fake profile zap attacks
Cc: Tony Giorgio <tonygiorgio@protonmail.com>
Cc: benthecarman <benthecarman@live.com>
Cc: Vitor Pamplona <vitor@vitorpamplona.com>
This fixes a zap issue where someone could send a fake zap with a zapper
that doesn't match the user's nostrPubkey zapper. This is possible
because damus looks up the zapper via the ptag on note zaps.
Fix this by first looking up the cached event's ptag instead. This
prevents zappers from trying to trick Damus into picking the wrong
zapper.
Fixes: #1357
Changelog-Fixed: Fix issue where malicious zappers can send fake zaps to another user's posts
Reported-by: benthecarman <benthecarman@live.com>
Cc: Tony Giorgio <tonygiorgio@protonmail.com>
