Skip to content

Latest commit

 

History

History
131 lines (85 loc) · 3.6 KB

README.md

File metadata and controls

131 lines (85 loc) · 3.6 KB

REALM with wolfSSL

The wolfSSL library is enabled in realm-core with the -DREALM_HAVE_WOLFSSL=1 parameter.

The upstream realm-core does NOT yet support wolfSSL. Please see the enclosed realm-commit-[hash].patch files.

The Linux Script Build automatically applies these patches as needed. Modify as desired with the REALM_CORE_COMMIT script variable.

Linux Script Build

The build_wolfssl_with_realm.sh script can be used.

Script Parameters:

  • -t use tarball, not git
  • -u use $USER name suffix for repository

Defaults are to clone wolfSSL and realm-core to local directory.

If you have forks of each with your GitHub ID the same as your local user name, the -u is recommended.

Run directly from the realm directory:

./build_wolfssl_with_realm.sh -u

This will create these local directories, and add respective upstream repository settings:

./wolfssl-<your username>
./realm-core-<your username>

Once the desired wolfSSL library has been fetched and installed, the clone / config / build can be skipped by making these changes in the script file:

# Choose to skip parts of wolfSSL build:
FETCH_WOLFSSL=false
CONFIGURE_WOLFSSL=false
BUILD_WOLFSSL=false
INSTALL_WOLFSSL=false

Linux manual build: Build wolfSSL

These commands are wrapped in the build_wolfssl_with_realm.sh script, mentioned above:

To configure wolfSSL, use the following command:

./configure --enable-static --enable-opensslall --enable-enckeys --enable-certgen --enable-context-extra-user-data
sudo make install

Configuring wolfSSL for local installation can be specified with --prefix=/path/to/install

Downloading and applying the patch for realm-core git commit a5e87a39:

git clone https://github.com/realm/realm-core.git
cd realm-core
git reset --hard HEAD
git checkout a5e87a39
git submodule update --init --recursive
git apply ../realm-v13.26.0.patch

Building realm-core:

mkdir build
cmake -B build -DREALM_ENABLE_ENCRYPTION=1 -DREALM_ENABLE_SYNC=1 -DREALM_HAVE_WOLFSSL=1 -DREALM_WOLFSSL_ROOT_DIR=/usr/local/lib
cmake --build build
./build/test/realm-tests

You can also use the build_wolfssl_with_realm.sh script after adjusting the global variables as needed.

Generating a new Realm-core patch file:

To generate a new patch compare a particular commit (a5e87a39) to your fork/branch (dev):

Save the generated file from this link to the realm-commit-a5e87a39.patch file:

https://github.com/gojimmypi/realm-core/compare/a5e87a39...dev.patch

or

https://github.com/gojimmypi/realm-core/compare/a5e87a39...gojimmypi:realm-core:dev-consolidated.patch

Do not edit the patch file for tailing spaces, etc. It must be saved exactly as generated.

Troubleshooting

Confirming build with wolfSSL

See generated config.h files:

For Linux compile: build directory:

[wolfssl osp root]\realm\realm-core\build\src\realm\util

For Visual Studio compile: out directory, plus build configuration (e.g. build\x64-Debug):

[wolfssl osp root]\realm\realm-core\out\build\x64-Debug\src\realm\util

Cannot find OpenSSL

Cannot open include file: 'openssl/sha.h': No such file or directory	Storage	C:\workspace\osp-gojimmypi\realm\realm-core-gojimmypi\src\realm\util\aes_cryptor.hpp

Check the osp\realm\VS2022\src\realm\util directory. The generated config.h should have openSSL disabled 0 and wolfSSL enabled 1 like this:

#define REALM_HAVE_OPENSSL 0
#define REALM_HAVE_WOLFSSL 1