Note
This manual contains information for developers working on GNU Taler and related components. It is not intended for a general audience.
Table of Contents
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.
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/.
Before you can obtain Git write access, you must sign the copyright agreement. As we collaborate closely with GNUnet, we use their copyright agreement – with the understanding that your contributions to GNU Taler are included in the assignment. You can find the agreement on the GNUnet site. Please sign and mail it to Christian Grothoff as he currently collects all the documents for GNUnet e.V.
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 (recommended: dev/$USER/...
),
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
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.
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).
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.
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.
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
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.
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
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.
The $HOME/envcfg.py
file contains (1) the name of the environment and (2) the version
of all components we build (in the form of a git rev).
The envcfg.py
for demo looks like this:
env = "demo"
tag = "demo-2019-10-05-01:
tag_gnunet = tag
tag_libmicrohttpd = tag
tag_exchange = tag
tag_merchant = tag
tag_bank = tag
tag_twister = tag
tag_landing = tag
tag_donations = tag
tag_blog = tag
tag_survey = tag
tag_backoffice = tag
tag_sync = tag
Currently only the variables env
and tag_${component}
are used.
When deploying to demo
, the envcfg.py
should be committed to deployment.git/envcfg/envcfg-demo-YYYY-MM-DD-SS.py
.
$ 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-prepare
$ taler-deployment-start
$ taler-deployment-arm -I # check everything works
# The following command sets up the 'blog' and 'donations' instances.
$ taler-config-instances
$ rm -rf ~/sources ~/local
$ git -C ~/deployment pull
$ cp ~/deployment/envcfg/$ENVCFGFILE ~/envcfg.py
$ taler-deployment bootstrap
$ taler-deployment build
$ taler-deployment-prepare
$ taler-deployment-restart
$ taler-deployment-arm -I # check everything works
As the demo
user, to switch to color ${COLOR}
,
run the following script from deployment/bin
:
$ taler-deployment switch-demo
GNU Taler uses a buildbot implementation (front end at https://buildbot.taler.net) to manage continuous integration. Buildbot documentation is at https://docs.buildbot.net/.
Here are some highlights:
buildbot-worker
command.Best Practices:
master.cfg
file, leave a comment specifying the server and user account that this WORKER is called from. (At this time, taler.net is the only server used by this implementation, but it’s still good practice.)buildbot-worker create-worker <workername> localhost <username> <password>
Then make sure there is a WORKER defined in master.cfg like: worker.Worker("<username>", "<password>")
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/
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/
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/
.
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/
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/
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/
The Postgres databases of the exchange and the auditor are versioned. See the 0000.sql file in the respective directory for documentation.
Every set of changes to the database schema must be stored in a new versioned SQL script. The scripts must have contiguous numbers. After any release (or version being deployed to a production or staging environment), existing scripts MUST be immutable.
Developers and operators MUST NOT make changes to database schema outside of this versioning.
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
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
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.
Set the version in configure.ac
. The commit being tagged should be
the change of the version.
Tag the current GANA version that works with the exchange and merchant and checkout that tag of gana.git (instead of master). Otherwise, if there are incompatible changes in GANA (like removed symbols), old builds could break.
Update the Texinfo documentation using the files from docs.git:
# Get the latest documentation repository
$ cd $GIT/docs
$ git pull
$ make texinfo
# The *.texi files are now in _build/texinfo
#
# This checks out the prebuilt branch in the prebuilt directory
$ git worktree add prebuilt prebuilt
$ cd prebuilt
# Copy the pre-built documentation into the prebuilt directory
$ cp -r ../_build/texinfo .
# Push and commit to branch
$ git commit -a -S -m "updating texinfo"
$ git status
# Verify that all files that should be tracked are tracked,
# new files will have to be added to the Makefile.am in
# exchange.git as well!
$ git push
# Remember $REVISION of commit
#
# Go to exchange
$ cd $GIT/exchange/doc/prebuilt
# Update submodule to point to latest commit
$ git checkout $REVISION
Finally, the Automake Makefile.am
files may have to be adjusted to
include new *.texi
files or images.
For bootstrap, you will need to install GNU Recutils.
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
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
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.
Our general setup is based on https://wiki.debian.org/DebianRepository/SetupWithReprepro
First, update at least the version of the Debian package in debian/changelog, and then run:
$ dpkg-buildpackage -rfakeroot -b -uc -us
in the respective source directory (GNUnet, exchange, merchant) to create the
.deb
files. Note that they will be created in the parent directory. This
can be done on gv.taler.net, or on another (secure) machine.
Next, the *.deb
files should be copied to gv.taler.net, say to
/root/incoming
. Then, run
# cd /var/www/repos/apt/debian/
# reprepro includedeb sid /root/incoming/*.deb
to import all Debian files from /root/incoming/
into the sid
distribution. If Debian packages were build against other distributions,
reprepro may need to be first configured for those and the import command
updated accordingly.
Finally, make sure to clean up /root/incoming/
(by deleting the
now imported *.deb
files).
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.
Internationalization (a.k.a “Translation”) is handled with Weblate (https://weblate.org) via our instance at https://weblate.taler.net/ .
At this time, this system is still very new for Taler.net and this documentation may be incorrect and is certainly incomplete.
At this time, anyone can register an account at https://weblate.taler.net/ to create translations. Registered users default to the Users and Viewers privilege level.
This is the breakdown of privilege levels in Weblate:
To upgrade from Users/Viewers, a superuser must manually augment your privileges. At this time, superusers are Christian, Florian, and Buck.
The GNU Taler project is probably the correct project for most Components and Translations falling under this guide. Please contact a superuser if you need another Project created.
Reference: https://docs.weblate.org/en/weblate-4.0.3/admin/projects.html#component-configuration
In Weblate, a Component is a subset of a Project and each Component contains N translations. A Component is generally associated with a Git repo.
To create a Component, log into https://weblate.taler.net/ with your Manager or higher credentials and choose + Add from the upper-right corner.
What follows is a sort of Wizard. You can find detailed docs at https://docs.weblate.org/. Here are some important notes about connecting your Component to the Taler Git repository:
Under https://weblate.taler.net/create/component/vcs/:
git+ssh://git@git.taler.net/<reponame>`
. Check with git remote -v
.git+ssh://git@git.taler.net/<reponame>`
Check with git remote -v
.https://git.taler.net/<repositoryname>.git/tree/{{filename}}?h={{branch}}#n{{line}}
where <repositoryname>
gets replaced but {{filename}}
and other items in braces are actual variables in the string.1 - Log into https://weblate.taler.net
2 - Navigate to Projects > Browse all projects
3 - Choose the Project you wish to contribute to.
4 - Choose the Component you wish to contribute to.
5 - Find the language you want to translate into. Click “Translate” on that line.
6 - Find a phrase and translate it.
You may also wish to refer to https://docs.weblate.org/ .
By default, our Weblate instance is set to accept translations in English, French, German, Italian, Russian, Spanish, and Portuguese. If you want to contribute a translation in a different language, navigate to the Component you want to translate for, and click “Start new translation” to begin. If you require a privilege upgrade, please contact a superuser with your request.
When asked, set the license to GPLv3 or later.
Set commit/push to manual only.
weblate.taler.net signs GPG commits with the GPG key CD33CE35801462FA5EB0B695F2664BF474BFE502, and the corresponding public key can be found at https://weblate.taler.net/keys/.
This means that contributions made through weblate will not be signed with the individual contributor’s key when they are checked into the Git repository, but with the weblate key.
There are currently three Android apps in the official Git repository:
Their git repositories are mirrored at Gitlab
to utilize their CI
and F-Droid’s Gitlab integration
to publish automatic nightly builds
for each change on the master
branch.
All three apps publish their builds to the same F-Droid nightly repository (which is stored as a git repository): https://gitlab.com/gnu-taler/fdroid-repo-nightly
You can download the APK files directly from that repository or add it to the F-Droid app for automatic updates by clicking the following link (on the phone that has F-Droid installed).
Note
Nightly apps can be installed alongside official releases and thus are meant only for testing purposes. Use at your own risk!
Note that this guide is different from other guides for building Android apps,
because it does not require you to run non-free software.
It uses the Merchant PoS Terminal as an example, but works as well for the other apps
if you replace merchant-terminal
with wallet
or cashier
.
First, ensure that you have the required dependencies installed:
Then you can get the app’s source code using git:
# Start by cloning the Android git repository
$ git clone https://git.taler.net/taler-android.git
# Change into the directory of the cloned repository
$ cd taler-android
# Find out which Android SDK version you will need
$ grep -i compileSdkVersion merchant-terminal/build.gradle
The last command will return something like compileSdkVersion 29
.
So visit the Android Rebuilds project
and look for that version of the Android SDK there.
If the SDK version is not yet available as a free rebuild,
you can try to lower the compileSdkVersion
in the app’s merchant-terminal/build.gradle
file.
Note that this might break things
or require you to also lower other versions such as targetSdkVersion
.
In our example, the version is 29
which is available,
so download the “SDK Platform” package of “Android 10.0.0 (API 29)”
and unpack it:
# Change into the directory that contains your downloaded SDK
$ cd $HOME
# Unpack/extract the Android SDK
$ unzip android-sdk_eng.10.0.0_r14_linux-x86.zip
# Tell the build system where to find the SDK
$ export ANDROID_SDK_ROOT="$HOME/android-sdk_eng.10.0.0_r14_linux-x86"
# Change into the directory of the cloned repository
$ cd taler-android
# Build the merchant-terminal app
$ ./gradlew :merchant-terminal:assembleRelease
If you get an error message complaining about build-tools
- > Failed to install the following Android SDK packages as some licences have not been accepted.
- build-tools;29.0.3 Android SDK Build-Tools 29.0.3
you can try changing the buildToolsVersion
in the app’s merchant-terminal/build.gradle
file
to the latest “Android SDK build tools” version supported by the Android Rebuilds project.
After the build finished successfully,
you will find your APK in merchant-terminal/build/outputs/apk/release/
.
Translations are managed with Taler’s weblate instance: https://weblate.taler.net/projects/gnu-taler/
To update translations, enter the taler-android git repository and ensure that the weblate remote exists:
$ git config -l | grep weblate
If it does not yet exist (empty output), you can add it like this:
$ git remote add weblate https://weblate.taler.net/git/gnu-taler/wallet-android/
Then you can merge in translations commit from the weblate remote:
# ensure you have latest version
$ git fetch weblate
# merge in translation commits
$ git merge weblate/master
Afterwards, build the entire project from source and test the UI to ensure that no erroneous translations (missing placeholders) are breaking things.
After extensive testing, the code making up a new release should get a signed git tag. The current tag format is:
$ git tag -s $APP-$VERSION
Nightly builds get published automatically (see above) after pushing code to the official repo. Actual releases get picked up by F-Droid’s official repository via git tags. So ensure that all releases get tagged properly.
Some information for F-Droid official repository debugging:
Google Play uploads are managed via Fastlane. Before proceeding, ensure that this is properly set up and that you have access to the Google Play API.
To release an app, enter into its respective folder and run fastlane:
$ bundle exec fastlane
Then select the deploy option. Note this requires access to the Google Play upload signing key set via the various environment variables in $app/fastlane/Fastfile.
All uploads are going to the beta track by default. These can be promoted to production later or immediately after upload if you feel daring.
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/ .
GNU Taler is developed primarily in C, Kotlin, Python and TypeScript.
These are the general coding style rules for Taler.
taler_
” (exception: platform.h),
and MUST live in src/include/taler_
” and
MUST NOT live in src/include/ and
SHOULD NOT be included from outside of their own directory-l LOGFILE
, its absence means write logs to stderrTALER_
, without subsystemnameTALER_
or any other prefixShell scripts should be avoided if at all possible. The only permissible uses of shell scripts in GNU Taler are:
./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
We so far have no specific guidelines, please follow best practices for the language.
Python code should be written and build against version 3.7 of Python.
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.
When using Python for writing small utilities, the following libraries are useful:
click
for argument parsing (should be preferred 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.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.
/* Without 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.
This section contains terminology that should be used and that should not be used in the user interface and help materials.
Refreshing is the internal technical terminology for the protocol to give change for partially spent coins
Use instead: “Obtaining change”
Coins are an internal construct, the user should never be aware that their balance is represented by coins if different denominations.
Use instead: “(Digital) Cash” or “(Wallet) Balance”
Has bad connotation of consumption.
Use instead: Customer or user.
The term used to describe the process of the merchant facilitating the download of the signed contract terms for an order.
Avoid. Generally events that relate to proposal downloads should not be shown to normal users, only developers. Instead, use “communication with mechant failed” if a proposed order can’t be downloaded.
Should be generally avoided, since Taler is only anonymous for the customer. Also some people are scared of anonymity (which as a term is also way too absolute, as anonymity is hardly ever perfect).
Use instead: “Privacy-preserving”, “Privacy-friendly”
The process of proving to the merchant that the customer is entitled to view a digital product again, as they already paid for it.
Use instead: In the event history, “re-activated digital content purchase” could be used. (FIXME: this is still not nice.)
Too ambiguous in the wallet.
Use instead: Purchase
The entity/service that gives out digital cash in exchange for some other means of payment.
In some contexts, using “Issuer” could also be appropriate. When showing a balance breakdown, we can say “100 Eur (issued by exchange.euro.taler.net)”. Sometimes we may also use the more generic term “Payment Service Provider” when the concept of an “Exchange” is still unclear to the reader.
This glossary is meant for developers. It contains some terms that we usually do not use when talking to end users or even system administrators.
accounting mechanism used by the exchange to track customer funds from incoming wire transfers. A reserve is created whenever a customer wires money to the exchange using a well-formed public key in the subject. The exchange then allows the customer’s wallet to withdraw up to the amount received in fresh coins from the reserve, thereby draining the reserve. If a reserve is not drained, the exchange eventually closes it.
Other definition: Funds set aside for future use; either the balance of a customer at the exchange ready for withdrawal, or the funds kept in the exchange;s bank account to cover obligations from coins in circulation.
This section describes various internal programs to make life easier for the developer.
taler-config-generate - tool to simplify Taler configuration generation
taler-config-generate [-C CURRENCY | ––currency=CURRENCY] [-c FILENAME | ––config=FILENAME] [-e | ––exchange] [-f AMOUNT | ––wirefee=AMOUNT] [-h | ––help] [-J JSON | ––wire-json-exchange=JSON] [-j JSON | ––wire-json-merchant=JSON] [-L LOGLEVEL | ––loglevel=LOGLEVEL] [-m | ––merchant] [-t | ––trusted] [-v | ––version] [-w WIREFORMAT | ––wire WIREFORMAT] [––bank-uri] [––exchange-bank-account] [––merchant-bank-account]
taler-config-generate can be used to generate configuration files for the Taler exchange or Taler merchants.