10. Developer’s Manual

Note

This manual contains information for developers working on GNU Taler and related components. It is not intended for a general audience.

10.3. Fundamentals

10.3.1. Bug Tracking

Bug tracking is done with Mantis (https://www.mantisbt.org/). The bug tracker is available at https://bugs.taler.net. A registration on the Web site is needed in order to use the bug tracker, only read access is granted without a login.

10.3.2. Code Repositories

Taler code is versioned with Git. For those users without write access, all the codebases are found at the following URL:

git://git.taler.net/<repository>

A complete list of all the existing repositories is currently found at https://git.taler.net/.

10.3.3. Committing code

To obtain Git access, you need to send us your SSH public key. Most core team members have administrative Git access, so simply contact whoever is your primary point of contact so far. You can find instructions on how to generate an SSH key in the Git book. If you have been granted write access, you first of all must change the URL of the respective repository to:

ssh://git@git.taler.net/<repository>

For an existing checkout, this can be done by editing the .git/config file.

The server is configured to reject all commits that have not been signed with GnuPG. If you do not yet have a GnuPG key, you must create one, as explained in the GNU Privacy Handbook. You do not need to share the respective public key with us to make commits. However, we recommend that you upload it to key servers, put it on your business card and personally meet with other GNU hackers to have it signed such that others can verify your commits later.

To sign all commits, you should run

$ git config --global commit.gpgsign true

You can also sign individual commits only by adding the -S option to the git commit command. If you accidentally already made commits but forgot to sign them, you can retroactively add signatures using:

$ git rebase -S

Whether you commit to a personal branch, a feature branch or to master should depend on your level of comfort and the nature of the change. As a general rule, the code in master must always build and tests should always pass, at least on your own system. However, we all make mistakes and you should expect to receive friendly reminders if your change did not live up to this simple standard. We plan to move to a system where the CI guarantees this invariant in the future.

In order to keep a linear and clean commits history, we advise to avoid merge commits and instead always rebase your changes before pushing to the master branch. If you commit and later find out that new commits were pushed, the following command will pull the new commits and rebase yours on top of them.

# -S instructs Git to (re)sign your commits
$ git pull --rebase -S

10.3.4. Observing changes

Every commit to the master branch of any of our public repositories (and almost all are public) is automatically sent to the gnunet-svn@gnu.org mailinglist. That list is for Git commits only, and must not be used for discussions. It also carries commits from our main dependencies, namely GNUnet and GNU libmicrohttpd. While it can be high volume, the lists is a good way to follow overall development.

10.3.5. Communication

We use the #taler channel on the Freenode IRC network and the taler@gnu.org public mailinglist for discussions. Not all developers are active on IRC, but all developers should probably subscribe to the low-volume Taler mailinglist. There are separate low-volume mailinglists for gnunet-developers (@gnu.org) and for libmicrohttpd (@gnu.org).

10.4. Taler Deployment on gv.taler.net

This section describes the GNU Taler deployment on gv.taler.net. gv is our server at BFH. It hosts the Git repositories, Web sites, CI and other services. Developers can receive an SSH account and e-mail alias for the system. As with Git, ask your primary team contact for shell access if you think you need it.

Our old server, tripwire, is currently offline and will likely go back online to host production systems for operating real Taler payments at BFH in the future.

10.4.1. DNS

DNS records for taler.net are controlled by the GNU Taler maintainers, specifically Christian and Florian. If you need a sub-domain to be added, please contact one of them.

10.4.2. User Acccounts

On gv.taler.net, there are four system users that are set up to serve Taler on the Internet:

  • taler-test: serves *.test.taler.net and gets automatically built by Buildbot.
  • taler-internal: serves *.int.taler.net, and does NOT get automatically built.

The following two users are never automatically built, and they both serve *.demo.taler.net. At any given time, only one is active and serves the HTTP requests from the outside; the other one can so be compiled without any downtime. If the compilation succeeds, the inactive user can be switched to become active (see next section), and vice versa.

  • demo-blue
  • demo-green

10.5. Demo Upgrade Procedure

Upgrading the demo environment should be done with care, and ideally be coordinated on the mailing list before. It is our goal for demo to always run a “working version” that is compatible with various published wallets.

Before deploying on demo, the same version of all components must be deployed and tested on int.

Please use the demo upgrade checklist to make sure everything is working.

10.5.1. Tagging components

All Taler components must be tagged with git before they are deployed on the demo environment, using a tag of the following form:

demo-YYYY-MM-DD-SS
YYYY = year
MM = month
DD = day
SS = serial

10.5.2. Environment Layout

Environments have the following layout:

$HOME/
  deployment (deployment.git checkout)
  envcfg.py  (configuration of the Taler environment)
  activate   (bash file, sourced to set environment variables)
  logs/      (log files)
  local/     (locally installed software)
  sources/   (sources repos of locally build components)
  sockets/   (unix domain sockets of running components)
  taler-data (on-disk state, public and private keys)
  .config/taler.conf (main Taler configuration file)

On demo-blue and demo-green, taler-data is a symlink pointing to $HOME/demo/shared-data instead of a directory.

10.5.3. Using envcfg.py

The $HOME/envcfg.py file contains (1) the name of the environment and (2) the version of all components we build (in the form of a git rev).

The envcfg.py for demo looks like this:

env = "demo"
tag = "demo-2019-10-05-01:
tag_gnunet = tag
tag_libmicrohttpd = tag
tag_exchange = tag
tag_merchant = tag
tag_bank = tag
tag_twister = tag
tag_landing = tag
tag_donations = tag
tag_blog = tag
tag_survey = tag
tag_backoffice = tag

Currently only the variables env and tag_${component} are used.

When deploying to demo, the envcfg.py should be committed to deployment.git/envcfg/envcfg-demo-YYYY-MM-DD-SS.py.

10.5.4. Bootstrapping an Environment

$ git clone https://git.taler.net/deployment.git ~/deployment
$ cp ~/deployment/envcfg/$ENVCFGFILE ~/envcfg.py
$ ./deployment/bin/taler-deployment bootstrap
$ source ~/activate
$ taler-deployment build
$ taler-deployment-config-generate
$ taler-deployment-keyup
$ taler-deployment-sign
$ taler-exchange-dbinit -r
$ taler-merchant-dbinit -r
$ taler-deployment-start

10.5.5. Upgrading an Existing Environment

$ rm -rf ~/sources ~/local
$ git -C ~/deployment pull
$ cp ~/deployment/envcfg/$ENVCFGFILE ~/envcfg.py
$ taler-deployment bootstrap
$ taler-deployment build
$ taler-deployment-keyup
$ taler-deployment-sign
$ taler-deployment-start

10.5.6. Switching Demo Colors

As the demo user, to switch to color ${COLOR}, run the following script from deployment/bin:

$ taler-deployment-switch-demo-${COLOR}

10.6. Environments and Builders on taler.net

10.6.1. Documentation Builder

All the Taler documentation is built by the user docbuilder that runs a Buildbot worker. The following commands set the docbuilder up, starting with a empty home directory.

# Log-in as the 'docbuilder' user.

$ cd $HOME
$ git clone git://git.taler.net/deployment
$ ./deployment/bootstrap-docbuilder

# If the previous step worked, the setup is
# complete and the Buildbot worker can be started.

$ buildbot-worker start worker/

10.6.2. Website Builder

Taler Websites, www.taler.net and stage.taler.net, are built by the user taler-websites by the means of a Buildbot worker. The following commands set the taler-websites up, starting with a empty home directory.

# Log-in as the 'taler-websites' user.

$ cd $HOME
$ git clone git://git.taler.net/deployment
$ ./deployment/bootstrap-sitesbuilder

# If the previous step worked, the setup is
# complete and the Buildbot worker can be started.

$ buildbot-worker start worker/

10.6.3. Code coverage

Code coverage tests are run by the lcovworker user, and are also driven by Buildbot.

# Log-in as the 'lcovworker' user.

$ cd $HOME
$ git clone git://git.taler.net/deployment
$ ./deployment/bootstrap-taler lcov

# If the previous step worked, the setup is
# complete and the Buildbot worker can be started.

$ buildbot-worker start worker/

The results are then published at https://lcov.taler.net/.

10.6.4. Service Checker

The user demo-checker runs periodic checks to see if all the *.demo.taler.net services are up and running. It is driven by Buildbot, and can be bootstrapped as follows.

# Log-in as the 'demo-checker' user

$ cd $HOME
$ git clone git://git.taler.net/deployment
$ ./deployment/bootstrap-demochecker

# If the previous step worked, the setup is
# complete and the Buildbot worker can be started.

$ buildbot-worker start worker/

10.6.5. Tipping reserve top-up

Both ‘test’ and ‘demo’ setups get their tip reserve topped up by a Buildbot worker. The following steps get the reserve topper prepared.

# Log-in as <env>-topper, with <env> being either 'test' or 'demo'

$ git clone git://git.taler.net/deployment
$ ./deployment/prepare-reservetopper <env>

# If the previous steps worked, then it should suffice to start
# the worker, with:

$ buildbot-worker start worker/

10.6.6. Producing auditor reports

Both ‘test’ and ‘demo’ setups get their auditor reports compiled by a Buildbot worker. The following steps get the reports compiler prepared.

# Log-in as <env>-auditor, with <env> being either 'test' or 'demo'

$ git clone git://git.taler.net/deployment
$ ./deployment/prepare-auditorreporter <env>

# If the previous steps worked, then it should suffice to start
# the worker, with:

$ buildbot-worker start worker/

10.7. Releases

10.7.1. Release Process and Checklists

Please use the release checklist

This document describes the process for releasing a new version of the various Taler components to the official GNU mirrors.

The following components are published on the GNU mirrors

  • taler-exchange (exchange.git)
  • taler-merchant (merchant.git)
  • talerdonations (donations.git)
  • talerblog (blog.git)
  • taler-bank (bank.git)
  • taler-wallet-webex (wallet-webex.git)

10.7.2. Tagging

Tag releases with an annotated commit, like

git tag -a v0.1.0 -m "Official release v0.1.0"
git push origin v0.1.0

10.7.3. Database for tests

For tests in the exchange and merchant to run, make sure that a database talercheck is accessible by $USER. Otherwise tests involving the database logic are skipped.

10.7.4. Exchange, merchant

Set the version in configure.ac. The commit being tagged should be the change of the version.

For the exchange test cases to pass, make install must be run first. Without it, test cases will fail because plugins can’t be located.

./bootstrap
./configure # add required options for your system
make dist
tar -xf taler-$COMPONENT-$VERSION.tar.gz
cd taler-$COMPONENT-$VERSION
make install check

10.7.5. Wallet WebExtension

The version of the wallet is in manifest.json. The version_name should be adjusted, and version should be increased independently on every upload to the WebStore.

./configure
make dist

10.7.6. Upload to GNU mirrors

See https://www.gnu.org/prep/maintain/maintain.html#Automated-FTP-Uploads

Directive file:

version: 1.2
directory: taler
filename: taler-exchange-0.1.0.tar.gz

Upload the files in binary mode to the ftp servers.

10.8. Continuous integration

CI is done with Buildbot (https://buildbot.net/), and builds are triggered by the means of Git hooks. The results are published at https://buildbot.taler.net/.

In order to avoid downtimes, CI uses a “blue/green” deployment technique. In detail, there are two users building code on the system, the “green” and the “blue” user; and at any given time, one is running Taler services and the other one is either building the code or waiting for that.

There is also the possibility to trigger builds manually, but this is only reserved to “admin” users.

10.9. Code Coverage

Code coverage is done with the Gcov / Lcov (http://ltp.sourceforge.net/coverage/lcov.php) combo, and it is run nightly (once a day) by a Buildbot worker. The coverage results are then published at https://lcov.taler.net/.

10.10. Coding Conventions

GNU Taler is developed primarily in C, Kotlin, Python and TypeScript.

10.10.1. Components written in C

These are the general coding style rules for Taler.

Naming conventions

  • include files (very similar to GNUnet):
    • if installed, must start with “taler_” (exception: platform.h), and MUST live in src/include/
    • if NOT installed, must NOT start with “taler_” and MUST NOT live in src/include/ and SHOULD NOT be included from outside of their own directory
    • end in “_lib” for “simple” libraries
    • end in “_plugin” for plugins
    • end in “_service” for libraries accessing a service, i.e. the exchange
  • binaries:
    • taler-exchange-xxx: exchange programs
    • taler-merchant-xxx: merchant programs (demos)
    • taler-wallet-xxx: wallet programs
    • plugins should be libtaler_plugin_xxx_yyy.so: plugin yyy for API xxx
    • libtalerxxx: library for API xxx
  • logging
    • tools use their full name in GNUNET_log_setup (i.e. ‘taler-exchange-keyup’) and log using plain ‘GNUNET_log’.
    • pure libraries (without associated service) use ‘GNUNET_log_from’ with the component set to their library name (without lib or ‘.so’), which should also be their directory name (i.e. ‘util’)
    • plugin libraries (without associated service) use ‘GNUNET_log_from’ with the component set to their type and plugin name (without lib or ‘.so’), which should also be their directory name (i.e. ‘exchangedb-postgres’)
    • libraries with associated service) use ‘GNUNET_log_from’ with the name of the service, which should also be their directory name (i.e. ‘exchange’)
  • configuration
    • same rules as for GNUnet
  • exported symbols
    • must start with TALER_[SUBSYSTEMNAME]_ where SUBSYSTEMNAME MUST match the subdirectory of src/ in which the symbol is defined
    • from libtalerutil start just with TALER_, without subsystemname
    • if scope is ONE binary and symbols are not in a shared library, use binary-specific prefix (such as TMH = taler-exchange-httpd) for globals, possibly followed by the subsystem (TMH_DB_xxx).
  • structs:
    • structs that are ‘packed’ and do not contain pointers and are thus suitable for hashing or similar operations are distinguished by adding a “P” at the end of the name. (NEW) Note that this convention does not hold for the GNUnet-structs (yet).
    • structs that are used with a purpose for signatures, additionally get an “S” at the end of the name.
  • private (library-internal) symbols (including structs and macros)
    • must not start with TALER_ or any other prefix
  • testcases
    • must be called “test_module-under-test_case-description.c”
  • performance tests
    • must be called “perf_module-under-test_case-description.c”

10.10.2. Shell Scripts

Shell scripts should be avoided if at all possible. The only permissible uses of shell scripts in GNU Taler are:

  • Trivial invocation of other commands.
  • Scripts for compatibility (e.g. ./configure) that must run on as many systems as possible.

When shell scripts are used, they MUST begin with the following set command:

# Make the shell fail on undefined variables and
# commands with non-zero exit status.
set -eu

10.10.3. Kotlin

We so far have no specific guidelines, please follow best practices for the language.

10.10.4. Python

Supported Python Versions

Python code should be written and build against version 3.7 of Python.

Style

We use yapf to reformat the code to conform to our style instructions. A reusable yapf style file can be found in build-common, which is intended to be used as a git submodule.

Python for Scripting

When using Python for writing small utilities, the following libraries are useful:

  • click for argument parsing (should be prefered over argparse)
  • pathlib for path manipulation (part of the standard library)
  • subprocess for “shelling out” to other programs. Prefer subprocess.run over the older APIs.

10.11. Testing library

This chapter is a VERY ABSTRACT description of how testing is implemented in Taler, and in NO WAY wants to substitute the reading of the actual source code by the user.

In Taler, a test case is a array of struct TALER_TESTING_Command, informally referred to as CMD, that is iteratively executed by the testing interpreter. This latter is transparently initiated by the testing library.

However, the developer does not have to defined CMDs manually, but rather call the proper constructor provided by the library. For example, if a CMD is supposed to test feature x, then the library would provide the TALER_TESTING_cmd_x () constructor for it. Obviously, each constructor has its own particular arguments that make sense to test x, and all constructor are thoroughly commented within the source code.

Internally, each CMD has two methods: run () and cleanup (). The former contains the main logic to test feature x, whereas the latter cleans the memory up after execution.

In a test life, each CMD needs some internal state, made by values it keeps in memory. Often, the test has to share those values with other CMDs: for example, CMD1 may create some key material and CMD2 needs this key material to encrypt data.

The offering of internal values from CMD1 to CMD2 is made by traits. A trait is a struct TALER_TESTING_Trait, and each CMD contains a array of traits, that it offers via the public trait interface to other commands. The definition and filling of such array happens transparently to the test developer.

For example, the following example shows how CMD2 takes an amount object offered by CMD1 via the trait interface.

Note: the main interpreter and the most part of CMDs and traits are hosted inside the exchange codebase, but nothing prevents the developer from implementing new CMDs and traits within other codebases.

/* Withouth loss of generality, let's consider the
 * following logic to exist inside the run() method of CMD1 */
..

struct TALER_Amount *a;
/**
 * the second argument (0) points to the first amount object offered,
 * in case multiple are available.
 */
if (GNUNET_OK != TALER_TESTING_get_trait_amount_obj (cmd2, 0, &a))
  return GNUNET_SYSERR;
...

use(a); /* 'a' points straight into the internal state of CMD2 */

In the Taler realm, there is also the possibility to alter the behaviour of supposedly well-behaved components. This is needed when, for example, we want the exchange to return some corrupted signature in order to check if the merchant backend detects it.

This alteration is accomplished by another service called twister. The twister acts as a proxy between service A and B, and can be programmed to tamper with the data exchanged by A and B.

Please refer to the Twister codebase (under the test directory) in order to see how to configure it.