iTowns needs nodejs
(10+) and npm
(at least 6.x) installed on the system.
See the node website to install node and npm.
- Search and use existing methods to avoid duplicate code
- Avoid instance
class/object
in method (optimization with the garbage collector) - Give explicite name to constants
- Avoid duplicate code and encapsulate to simplify
- For performance reasons, update objects only when needed
- Don't add dead code
- Remove obsolete comments
- Download dependencies:
npm install
- Run the dev server:
npm start
(you can change the port:npm start -- --port 3000
. If you get an error message that is not related to the port being in use, please update to last lts or stable version)- Open
http://localhost:8080/
- Make changes in the code, the browser will automatically reload on save
- Build iTowns to produce a single independent script:
npm run build
- The script and its source-maps are generated into the
dist/
folder - You can run the built script to verify it works OK:
python -m SimpleHTTPServer 8080
(change the port at your convenience)- Open
http://localhost:8080/
- Transpile itowns to ES5 to require it in your code:
npm run transpile
- The transpiled source is in
lib/
- You can then require
src/Main.js
in your code
- Test and lint changes: see test, lint and test-examples npm script
babel-inline-import-loader
prevents the source map debug in browser. If you want launch server and debug with the original source map, run :npm run debug
.- To debug iTowns package on your side project. You can link iTowns package with
npm link ../path/iTowns/
in project folder and auto-transpile tolib/
when iTowns sources are modified with commandnpm run watch
in iTowns folder.
For unit and functional test, defines HTTPS_PROXY
if you launch test behind a proxy.
In order to run the tests, puppeteer
needs to be installed. If installing puppeteer behind a proxy, use HTTP_PROXY
,
HTTPS_PROXY
, NO_PROXY
to defines HTTP proxy settings that are used to download
and run Chromium.
If puppeteer fails to download Chrome, you can try with the documented environment variables. Or you can download it manually, and then:
- install puppeteer without downloading Chrome:
PUPPETEER_SKIP_CHROMIUM_DOWNLOAD=1 npm install puppeteer
- then use the env var
CHROME
to tell itowns/mocha/puppeteer what Chrome app it should use:CHROME=/opt/google/chrome-beta/chrome npm run test-examples
Then tests can be run with five differents methods:
npm run test
: build and run all tests in iTowns (the one used by github action)npm run test-dev
: build in development mode and run all tests in iTowns (to get more messages)npm run test-unit
: run unit tests onlynpm run test-functional
: run functional testing with examples only, usenpx mocha -t 30000 test/functional/bootstrap.js test/functional/<test_case>.js
to run a single examplenpm run test-with-coverage
: build and run all tests in iTowns and generate a report on the coverage of the tests
When running tests on examples, some environment variables can be set:
SCREENSHOT_FOLDER
: take a screenshot at the end of each test and save it in this folder. Example:SCREENSHOT_FOLDER=/tmp/
CHROME
: path to Chrome executable. If unspecified itowns will use the one downloaded during puppeteer install.DEBUG
: run Chrome in a window with the debug tools open.REMOTE_DEBUGGING
: run Chrome in headless mode and set up remote debugging. Example:REMOTE_DEBUGGING=9222
will setup remote debugging on port 9222. Then start another Chrome instance, browse tochrome://inspect/#devices
and addlocalhost:9222
in Discover network targets.
Note: Chrome in headless mode doesn't support the WebGL EXT_frag_depth
extension. So rendering may differ and some bugs can only be present in headless
mode.
mochajs is used for both type of tests. To avoid
problems with same name variables, keep them in the smallest scope. For example,
a variable should almost always be in the it()
section of a test. However, it
can be useful to keep a single shared variable for a bunch of tests (in the same
file). For this, declare it in the describe()
section, and set it (if
possible) in a before()
section.
If you wish to have unit tests to continuously run, you can append to your
running command -- --watch
. You will be getting in this case npm run test-unit -- --watch
.
If you want to work on a single test and debug it, without having all the extra
output, you can use this command npm run base-test-unit test/unit/3dtiles.js
and of course replace 3dtiles.js with the correct filename. You can also append
the -- --watch
flag as well. If you want disable reporting and keep error message
in console add --reporter min
.
npm run changelog
generates changelog file.npm run bump --level=minor
bump version, generates changelog file and commit (default level is minor).
See the contributor's guide for more information on how to contribute.
Note : You should not commit changes to package-lock.json
if you're not using
npm >= 6 (particularly if you didn't make any change to the package.json
either).