Skip to content

Latest commit

 

History

History
641 lines (470 loc) · 25 KB

README_maintainers.md

File metadata and controls

641 lines (470 loc) · 25 KB

ArangoDB Maintainers manual

This file contains documentation about the build process, documentation generation means, unittests - put short - if you want to hack parts of arangod this could be interesting for you.

Configure

  • --enable-relative - relative mode so you can run without make install
  • --enable-maintainer-mode - generate lex/yacc files
  • --with-backtrace - add backtraces to native code asserts & exceptions
  • --enable-failure-tests - adds javascript hook to crash the server for data integrity tests

CFLAGS

Add backtraces to cluster requests so you can easily track their origin:

-DDEBUG_CLUSTER_COMM

V8 Special flags:

-DENABLE_GDB_JIT_INTERFACE

(enable (broken) GDB intergation of JIT) At runtime arangod needs to be started with these options:

--javascript.v8-options="--gdbjit_dump"
--javascript.v8-options="--gdbjit_full"

Debugging the Make process

If the compile goes wrong for no particular reason, appending 'verbose=' adds more output. For some reason V8 has VERBOSE=1 for the same effect.

Runtime

  • start arangod with --console to get a debug console
  • Cheapen startup for valgrind: --no-server --javascript.gc-frequency 1000000 --javascript.gc-interval 65536 --scheduler.threads=1 --javascript.v8-contexts=1
  • to have backtraces output set this on the prompt: ENABLE_NATIVE_BACKTRACES(true)

Startup

We now have a startup rc file: ~/.arangod.rc. It's evaled as javascript. A sample version to help working with the arangod rescue console may look like that:

ENABLE_NATIVE_BACKTRACES(true);
internal = require("internal");
fs = require("fs");
db = internal.db;
time = internal.time;
timed = function (cb) {
  var s  = time();
  cb();
  return time() - s;
};
print = internal.print;

JSLint

(we switched to jshint a while back - this is still named jslint for historical reasons)

Make target

use

make gitjslint

to lint your modified files.

make jslint

to find out whether all of your files comply to jslint. This is required to make continuous integration work smoothly.

if you want to add new files / patterns to this make target, edit js/Makefile.files

to be safe from committing non-linted stuff add .git/hooks/pre-commit with:

make gitjslint

Use jslint standalone for your js file

If you want to search errors in your js file, jslint is very handy - like a compiler is for C/C++. You can invoke it like this:

bin/arangosh --jslint js/server/modules/@arangodb/testing.js

ArangoDB Unittesting Framework

Dependencies

  • Ruby, rspec, httparty, boost_test (compile time)

Filename conventions

Special patterns in the test filenames are used to select tests to be executed or skipped depending on parameters:

-cluster

These tests will only run if clustering is tested. (option 'cluster' needs to be true).

-noncluster

These tests will only run if no cluster is used. (option 'cluster' needs to be false)

-timecritical

These tests are critical to execution time - and thus may fail if arangod is to slow. This may happen i.e. if you run the tests in valgrind, so you want to avoid them since they will fail anyways. To skip them, set the option skipTimeCritical to true.

-disabled

These tests are disabled. You may however want to run them by hand.

replication

(only applies to ruby tests) These tests aren't run automatically since they require a manual set up environment.

-spec

These tests are ran using the jasmine framework instead of jsunity.

-nightly

These tests produce a certain thread on infrastructure or the test system, and therefore should only be executed once per day.

Test frameworks used

There are several major places where unittests live:

  • UnitTests/HttpInterface - rspec tests
  • UnitTests/Basics - boost unittests
  • UnitTests/Geo - boost unittests
  • js/server/tests - runneable on the server
  • js/common/tests - runneable on the server & via arangosh
  • js/common/test-data
  • js/client/tests - runneable via arangosh
  • js/apps/system/aardvark/test

HttpInterface - RSpec Client Tests

These tests work on the plain RESTfull interface of arangodb, and thus also test invalid HTTP-requests and so forth, plus check error handling in the server.

jsUnity on arangod

you can engage single tests when running arangod with console like this:

require("jsunity").runTest("js/server/tests/aql-queries-simple.js");

jsUnity via arangosh

arangosh is similar, however, you can only run tests which are intended to be ran via arangosh:

require("jsunity").runTest("js/client/tests/shell-client.js");

jasmine tests

Jasmine tests cover two important usecase:

  • testing the UI components of aardvark -spec

aardvark

Invocation methods

Make-targets

Most of the tests can be invoked via the main Makefile: (UnitTests/Makefile.unittests)

  • unittests
  • unittests-brief
  • unittests-verbose
  • unittests-recovery
  • unittests-config
  • unittests-boost
  • unittests-single
  • unittests-shell-server
  • unittests-shell-server-only
  • unittests-shell-server-aql
  • unittests-shell-client-readonly
  • unittests-shell-client
  • unittests-http-server
  • unittests-ssl-server
  • unittests-import
  • unittests-replication
  • unittests-replication-server
  • unittests-replication-http
  • unittests-replication-data
  • unittests-upgrade
  • unittests-dfdb
  • unittests-foxx-manager
  • unittests-dump
  • unittests-arangob
  • unittests-authentication
  • unittests-authentication-parameters

Javascript framework

(used in our local Jenkins and TravisCI integration; required for running cluster tests)

Invoked like that:

scripts/unittest all

calling it without parameters like this:

scripts/unittest

will give you a extensive usage help which we won't duplicate here.

Choosing facility


The first parameter chooses the facility to execute. Available choices include:

  • all: (calls multiple) This target is utilized by most of the jenkins builds invoking unit tests.
  • single_client: (see Running a single unittestsuite)
  • single_server: (see Running a single unittestsuite)
  • single_localserver: (see Running a single unittestsuite)
  • many more - call without arguments for more details.

Passing Options

Different facilities may take different options. The above mentioned usage output contains the full detail.

A commandline for running a single test (-> with the facility 'single_server') using valgrind could look like this. Options are passed as regular long values in the syntax --option value --sub:option value. Using Valgrind could look like this:

./scripts/unittest single_server --test js/server/tests/aql-escaping.js \
  --extraargs:server.threads 1 \
  --extraargs:scheduler.threads 1 \
  --extraargs:javascript.gc-frequency 1000000 \
  --extraargs:javascript.gc-interval 65536 \
  --valgrind /usr/bin/valgrind \
  --valgrindargs:log-file /tmp/valgrindlog.%p
  • we specify the test to execute
  • we specify some arangod arguments via --extraargs which increase the server performance
  • we specify to run using valgrind (this is supported by all facilities
  • we specify some valgrind commandline arguments

Running a single unittestsuite

Testing a single test with the framework directly on a server:

scripts/unittest single_server --test js/server/tests/aql-escaping.js

Testing a single test with the framework via arangosh:

scripts/unittest single_client --test js/server/tests/aql-escaping.js

Testing a single rspec test:

scripts/unittest http_server --test api-users-spec.rb

scripts/unittest is mostly only a wrapper; The backend functionality lives in: js/server/modules/@arangodb/testing.js

Running foxx tests with a fake foxx Repo

Since downloading fox apps from github can be cumbersome with shaky DSL and DOS'ed github, we can fake it like this:

export FOXX_BASE_URL="http://germany/fakegit/"
./scripts/unittest single_server --test 'js/server/tests/shell-foxx-manager-spec.js'

arangod Emergency console

require("jsunity").runTest("js/server/tests/aql-escaping.js");

arangosh client

require("jsunity").runTest("js/server/tests/aql-escaping.js");

arangod commandline arguments

bin/arangod /tmp/dataUT --javascript.unit-tests="js/server/tests/aql-escaping.js" --no-server

make unittest

js/common/modules/loadtestrunner.js

Windows debugging

For the average *nix user windows debugging has some awkward methods.

Coredump generation

Coredumps can be created using the task manager; switch it to detail view, the context menu offers to create dump file; the generated file ends in a directory that explorer hides from you - AppData - you have to type that in the location bar. This however only for running processes which is not as useful as having dumps of crashing processes. While its a common feature to turn on coredumps with the system facilities on *nix systems, its not as easy in windows. You need an external program from the Sysinternals package: ProcDump. First look up the PID of arangod, you can finde it in the brackets in the arangodb logfile. Then invoke procdump like this:

procdump -accepteula -e -ma < PID of arangod >

It will keep on running and monitor arangod until eventually a crash happens. You will then get a core dump if an incident occurs or Dump count not reached. if nothing happened, Dump count reached. if a dump was written - the filename will be printed above.

Debugging symbols

Releases are supported by a public symbol server so you will be able to debug cores. Releases starting with 2.5.6, 2.6.3 onwards are supported; Note that you should run the latest version of a release series before reporting bugs. Either WinDbg or Visual studio support setting the symbol path via the environment variable or in the menu. Given we want to store the symbols on e:\symbol_cach we add the arangodb symbolserver like this:

set _NT_SYMBOL_PATH=cache*e:\symbol_cache\cache;srv*e:\symbol_cache\arango*https://www.arangodb.com/repositories/symsrv/;SRV*e:\symbol_cache\ms*http://msdl.microsoft.com/download/symbols

You then will be able to see stack traces in the debugger.

Coredump analysis

While Visual studio may cary a nice shiny gui, the concept of GUI fails miserably i.e. in testautomation. Getting an overview over all running threads is a tedious task with it. Here the commandline version of WinDBG cdb comes to the aid. testing.js utilizes it to obtain automatical stack traces for crashes. We run it like that:

cdb -z <dump file> -c 'kp; ~*kb; dv; !analyze -v; q'

These commands for -c mean: kp print curren threads backtrace with arguments ~*kb print all threads stack traces dv analyze local variables (if) !analyze -v print verbose analysis q quit the debugger

If you don't specify them via -c you can also use them in an interactive manner.


Documentation

Dependencies to build documentation:

  • swagger 2 for the API-Documentation inside aardvark (no installation required)

  • Python Setuptools

    https://pypi.python.org/pypi/setuptools

    Download setuptools zip file, extract to any folder, use bundled python 2.6 to install:

      python ez_install.py
    

    This will place the required files in python's Lib/site-packages folder.

  • MarkdownPP

    https://github.com/triAGENS/markdown-pp/

    Checkout the code with Git, use bundled python 2.6 to install:

      python setup.py install
    
  • Node.js

    Make sure the option to Add to PATH is enabled. After installation, the following commands should be available everywhere:

    • node
    • npm

    If not, add the installation path to your environment variable PATH.

  • Gitbook

    Can be installed with Node's package manager NPM:

      npm install gitbook-cli -g
    
  • Calibre2 (optional, only required if you want to build the e-book version)

    http://calibre-ebook.com/download

    Run the installer and follow the instructions.

Generate users documentation

If you've edited REST-Documentation, first invoke make swagger. If you've edited examples, see below howto regenerate them. Run the make command in arangodb/Documentation/Books to generate it. The documentation will be generated into arangodb/Documentation/Books/books/Users - use your favourite browser to read it.

You may encounter permission problem with gitbook and its NPM invokations; In that case you need to run the command as root / Administrator.

On windows you may see "device busy" errors, retry. Make shure you don't have intermediate files in the ppbooks / books -sub folder open (i.e. browser or editor) It can also temporarily occur during phases of high HDD / SSD load.

The build-scripts contain several sanity checks, i.e. whether all examples are used, and no dead references are there. (see building examples in that case below)

If the markdown files aren't converted to html, or index.html shows a single chapter only (content of README.md), make sure Cygwin create native symlinks It does not, if SUMMARY.md in Books/ppbooks/ looks like this:

!<symlink>ÿþf o o

If sub-chapters do not show in the navigation, try another browser (Firefox). Chrome's security policies are pretty strict about localhost and file:// protocol. You may access the docs through a local web server to lift the restrictions.

To only regereneate one file (faster) you may specify a filter:

make FILTER=Users/Aql/Invoke.mdpp

(regular expressions allowed)

Using Gitbook

The make command in arangodb/Documentation/Books/ generates a website version of the manual. If you want to generate PDF, ePUB etc., run below build commands in arangodb/Documentation/Books/books/Users/. Calibre's ebook-convert will be used for the conversion.

Generate a PDF:

gitbook pdf ./ppbooks/Users ./target/path/filename.pdf

Generate an ePub:

gitbook epub ./ppbooks/Users ./target/path/filename.epub

Where to add new...

  • js/action/api/* - markdown comments in source with execution section
  • Documentation/Books/Users/SUMMARY.md - index of all sub documentations
  • Documentation/Scripts/generateSwaggerApi.py - list of all sections to be adjusted if

generate

  • ./scripts/generateExamples --onlyThisOne geoIndexSelect will only produce one example - geoIndexSelect
  • ./scripts/generateExamples --onlyThisOne 'MOD.*' will only produce the examples matching that regex; Note that examples with enumerations in their name may base on others in their series - so you should generate the whole group.
  • ./scripts/generateExamples --server.endpoint tcp://127.0.0.1:8529 will utilize an existing arangod instead of starting a new one. this does seriously cut down the execution time.
  • alternatively you can use generateExamples (i.e. on windows since the make target is not portable) like that: ./scripts/generateExamples --server.endpoint 'tcp://127.0.0.1:8529' --withPython 3rdParty/V8-4.3.61/third_party/python_26/python26.exe --onlyThisOne 'MOD.*'
  • ./Documentation/Scripts/allExamples.sh generates a file where you can inspect all examples for readability.
  • make swagger - on top level to generate the documentation interactively with the server; you may use the swagger editor to revalidate whether js/apps/system/_admin/aardvark/APP/api-docs.json is accurate.
  • cd Documentation/Books; make - to generate the HTML documentation

write markdown

mdpp files are used for the structure. To join it with parts extracted from the program documentation you need to place hooks:

  • @startDocuBlock <tdocuBlockName> is replaced by a Docublock extracted from source.
  • @startDocuBlockInline <docuBlockName> till @endDocuBlock <docuBlockName> is replaced in with its own evaluated content - so @EXAMPLE_ARANGOSH_[OUTPUT | RUN] sections are executed the same way as inside of source code documentation.

Read / use the documentation

  • file:///Documentation/Books/books/Users/index.html contains the generated documentation
  • JS-Console - Tools/API - Interactive swagger documentation which you can play with.

arangod Example tool

make example picks examples from the source code documentation, executes them, and creates a transcript including their results. Hint: Windows users may use ./scripts/generateExamples for this purpose

Here is how its details work:

  • all ending with .cpp, .js and .mdpp are searched.
  • all lines inside of source code starting with '///' are matched, all lines in .mdpp files.
  • an example start is marked with @EXAMPLE_ARANGOSH_OUTPUT or @EXAMPLE_ARANGOSH_RUN
  • the example is named by the string provided in brackets after the above key
  • the output is written to Documentation/Examples/<name>.generated
  • examples end with @END_EXAMPLE_[OUTPUT|RUN]
  • all code in between is executed as javascript in the arangosh while talking to a valid arangod. You may inspect the generated js code in /tmp/arangosh.examples.js

OUTPUT and RUN specifics

By default, Examples should be self contained and thus not depend on each other. They should clean up the collections they create. Building will fail if resources aren't cleaned. However, if you intend a set of OUTPUT and RUN to demonstrate interactively and share generated ids, you have to use an alphabetical sortable naming scheme so they're executed in sequence. Using <modulename>_<sequencenumber>[a|b|c|d]_thisDoesThat seems to be a good scheme.

  • OUTPUT is intended to create samples that the users can cut'n'paste into their arangosh. Its used for javascript api documentation.

    • wrapped lines: Lines starting with a pipe (/// |) are joined together with the next following line. You have to use this if you want to execute loops, functions or commands which shouldn't be torn apart by the framework.
    • Lines starting with var: The command behaves similar to the arangosh: the server reply won't be printed. However, the variable will be in the scope of the other lines - else it won't.
    • Lines starting with a Tilde (/// ~): These lines can be used for setup/teardown purposes. They are completely invisible in the generated example transcript.
    • ~removeIgnoreCollection("test") - the collection test may live across several tests.
    • ~addIgnoreCollection("test") - the collection test will again be alarmed if left over.
  • it is executed line by line. If a line is intended to fail (aka throw an exception), you have to specify // xpError(ERROR_ARANGO_DOCUMENT_KEY_UNEXPECTED) so the exception will be caught; else the example is marked as broken. If you need to wrap that line, you may want to make the next line starting by a tilde to suppress an empty line:

    /// | someLongStatement()
    /// ~ // xpError(ERROR_ARANGO_DOCUMENT_KEY_UNEXPECTED)
    
  • RUN is intended to be pasted into a unix shell with cURL to demonstrate how the REST-HTTP-APIs work. The whole chunk of code is executed at once, and is expected not to throw. You should use assert(condition) to ensure the result is what you've expected. The body can be a string, or a javascript object which is then represented in JSON format.

    • Send the HTTP-request: var response = logCurlRequest('POST', url, body);
    • check its response: assert(response.code === 200);
    • output a JSON server Reply: logJsonResponse(response); (will fail if its not a valid json)
    • output HTML to the user: logHtmlResponse(response); (i.e. redirects have HTML documents)
    • output the plain text to dump to the user: logRawResponse(response);
    • dump the reply to the errorlog for testing (will mark run as failed): logErrorResponse(response);

Swagger integration

make swagger scans the sourcecode, and generates swagger output. It scans for all documentationblocks containing @RESTHEADER. It is a prerequisite for integrating these blocks into the gitbook documentation.

Tokens may have curly brackets with comma separated fields. They may optionally be followed by subsequent text lines with a long descriptions.

Sections group a set of tokens; they don't have parameters.

Types Swagger has several native types referenced below: [integer|long|float|double|string|byte|binary|boolean|date|dateTime|password] - see the swagger documentation

It consists of several sections which in term have sub-parameters:

Supported Sections:

RESTQUERYPARAMETERS

Parameters to be appended to the URL in form of ?foo=bar add RESTQUERYPARAMs below

RESTURLPARAMETERS

Section of parameters placed in the URL in curly brackets, add RESTURLPARAMs below.

RESTDESCRIPTION

Long text describing this route.

RESTRETURNCODES

should consist of several RESTRETURNCODE tokens.

Supported Tokens:

RESTHEADER

Up to 2 parameters.

  • [GET|POST|PUT|DELETE] url should start with a /, it may contain parameters in curly brackets, that have to be documented in subsequent RESTURLPARAM tokens in the RESTURLPARAMETERS section.

RESTURLPARAM

Consists of 3 values separated by ',': Attributes:

  • name: name of the parameter
  • type:
  • [required|optionas] Optional is not supported anymore. Split the docublock into one with and one without.

Folowed by a long description.

RESTALLBODYPARAM

This API has a schemaless json body (in doubt just plain ascii)

RESTBODYPARAM

Attributes:

  • name - the name of the parameter
  • type - the swaggertype of the parameter
  • required/optional - whether the user can ommit this parameter
  • subtype / format (can be empty)
    • subtype: if type is object or array, this references the enclosed variables. can be either a swaggertype, or a RESTRUCT
    • format: if type is a native swagger type, some support a format to specify them

RESTSTRUCT

Groups a set of attributes under the structure name to an object that can be referenced in other RESTSTRUCTs or RESTBODYPARAM attributes of type array or object

Attributes:

  • name - the name of the parameter
  • structure name - the type under which this structure can be reached (should be uniq!)
  • type - the swaggertype of the parameter (or another RESTSTRUCT...)
  • required/optional - whether the user can ommit this parameter
  • subtype / format (can be empty)
    • subtype: if type is object or array, this references the enclosed variables. can be either a swaggertype, or a RESTRUCT
    • format: if type is a native swagger type, some support a format to specify them

Local cluster startup

There are two scripts scripts/startLocalCluster and scripts/stopLocalCluster which help you to quickly fire up a testing cluster on your local machine. scripts/startLocalCluster takes 0, 2 or three arguments. In the 0 argument version, it simply starts 2 DBservers and one coordinator in the background, running on ports 8629, 8630 and 8530 respectively. The agency runs on port 4001. With 2 arguments the first is the number of DBServers and the second is the number of coordinators.

If there is a third argument and it is "C", then the first coordinator will be started with --console`` in a separate window (using an xterm`).

If there is a third argument and it is "D", then all servers are started up in the GNU debugger in separate windows (using xterms). In that case one has to hit ENTER in the original terminal where the script runs to continue, once all processes have been start up in the debugger.