status-im / nim-sqlcipher

SQLCipher Wrapper

nim-sqlcipher

Nim wrapper for SQLCipher. It builds SQLCipher and provides a simple API based on tiny_sqlite.

Requirements

# Linux
sudo apt install libssl-dev

# MacOS
brew install openssl

# Windows (msys2)
pacman -S mingw-w64-x86_64-openssl

Usage

TODO

License

Wrapper License

Licensed and distributed under the MIT License.

Dependency Licenses

• OpenSSL https://github.com/openssl/openssl/blob/master/LICENSE.txt
• SQLCipher https://github.com/sqlcipher/sqlcipher/blob/master/LICENSE
• nim-stew https://github.com/status-im/nim-stew#license
• nimbus-build-system https://github.com/status-im/nimbus-build-system/#license
• tiny_sqlite https://github.com/GULPF/tiny_sqlite/blob/master/LICENSE

nim-sqlcipher's Issues

Reconcile migration from nimterop to embedded tiny_sqlite with lates changes to this repo's master branch

Involves rebasing (and refactoring a bit) #18.

feat: prevent SQL injection attacks

As suggested in the nim manual, we can prevent SQL injection attacks by requiring a distinct type of string for DbValue.strVal.

tiny_sqlite does not support automatic sql injection attacks on string types, as outlined in the open issue: GULPF/tiny_sqlite#11.

Possible solution: https://nim-lang.org/docs/db_common.html#sql.t%2Cstring

Successfully build for Windows in a local msys2 environment

PR #12 (which is really great!) uses a proper msys2 environment for Windows instead of the minimal one provided by Git for Windows.

Before merging that PR, I want/need to become a bit more familiar with setting up a local msys2 environment. Also, nim-status' GitHub Actions workflow should be modified to work in the same way as it's refactored in #12 for this repo. Once that's done then #12 can be merged along with the corresponding PR in nim-status.

Try to figure out why TRUE/FALSE literals aren't working as expected

Related to #18 and status-im/nim-status#113.

SQLite does support use of TRUE and FALSE literals, translating them automatically to 1 and 0, respectively, since it doesn't have a boolean type per se.

See: https://www.db-fiddle.com/f/6RiUjLadVjuoh8otChkagD/0

However, that's not what we're experiencing, requiring us to use 1/0 instead of TRUE/FALSE. That's really not ideal and prone to humans forgetting (and writing bugs) as we e.g. use this library within nim-status.

Update the README

This repo's README needs to be updated with instructions and explanations stemming from the changes introduced in PRs #2 and #6.