Gdbc Save Abandoned

JDBC driver wrapper for golang

Project README

GDBC - Golang JDBC Driver Wrapper

CircleCI Godoc license

Load and use JDBC Java database drivers from go. Without Java.

IMPORTANT NOTES:

  • This is VERY MUCH WIP - it 'works' but has barely been tested/benchmarked. PRs/bug reports/feature requests gratefully received.
  • This repository will change. Do not rely on it yet. Compatibility is not guaranteed.
  • Very likely the drivers will be split into their own modules or repos.

Supported Drivers

More may be supported, but these are the ones we package, test, and distribute.

  • Microsoft SQL Server (mssql)
    • com.microsoft.sqlserver:mssql-jdbc:7.3.0.jre8-preview
  • Oracle DB (oracle)
    • com.oracle:ojdbc6:12.1.0.1 (NOT distributed, as license appears to prohibit it)
  • PostgreSQL (postgresql)
    • org.postgresql:postgresql:42.2.5.jre6

Building

For oracle, or if you want to build the native libraries yourself...

On osx run ./build-all.sh to build for both mac and linux.

On linux, run individual scripts e.g. ./wrapper/scripts/wrap-oracle.sh

Usage

Import oracle, mssql, or postgresql driver.

import _ "github.com/identitii/gdbc/oracle"

Use gdbc-(oracle|mssql|postgresql) as the driver name, and pass in the JDBC connection string (which is passed straight through to the driver)

pool, err = sql.Open("gdbc-oracle", "jdbc:oracle:thin:user/password@host:1521:sid")

When starting your application, it will look for the compiled driver shared library in the current directory (@executable_path) e.g. libgdbc-postgresql.dylib on mac, or libgdbc-postgresql.so on linux.

So either put the file in the current directory, or tell it where to find it like this...

DYLD_LIBRARY_PATH="path/to/the/shared/library" go run main.go

Benchmarks

The one tiny benchmark so far has these GDBC drivers about 30% faster than the pure go drivers on mac, and about 30% slower than the pure go drivers on linux.

How is this done?

GraalVM to package the java driver as a native library, then CGO to communicate with it.

Why is this done?

Because there are some databases that don't have or have incomplete or unsupported go drivers.

Initially started in order to get a fully featured oracle driver.

Also because it is fun.

TODO

  • Find trusted maven repositories to download drivers
  • Transaction isolation level config via driver
  • Distribute binary drivers (.dylib and .so, and .dll?) via separate repositories
  • Cut down the binary size from 120mb to something slightly more reasonable. (done: now 30mb)
    • Get rid of -H:IncludeResources=".*"
  • Extend test suite, find existing suite?
  • Test memory usage/leaking
  • Linux builds
  • Windows builds (waiting on graalvm support)
  • CI/CD
  • Implement further interfaces
    • sql/driver.Pinger
    • sql/driver.RowsColumnTypeDatabaseTypeName
    • sql/driver.RowsColumnTypeLength
    • sql/driver.RowsColumnTypeNullable
    • sql/driver.RowsColumnTypePrecisionScale
    • sql/driver.NamedValueChecker
  • Allow driver specific extensions

Drivers

  • sqlite? ("org.sqlite.JDBC" "jdbc:sqlite::memory:" "SELECT 1;")

Who Created This

We're a fintech startup based in Sydney, Australia.

Find this project interesting? Why not come work with us.

https://identitii.com/careers/

Thanks

https://github.com/japettyjohn/go-jdbc, which itself is partly based on https://github.com/xoba/goutil

gopher logo by Mario Furmanczyk (mfurmanc on fiverr)

Open Source Agenda is not affiliated with "Gdbc" Project. README Source: identitii/gdbc
Stars
98
Open Issues
5
Last Commit
1 year ago
License
MIT

Open Source Agenda Badge

Open Source Agenda Rating