Homec4science
Diffusion Phabricator (feature/testing) Browse

phab/src/docs/contributor/developer_setup.diviner
cef2e9b5c83cfeature/testing

Last Change
Raw File

developer_setup.diviner
View Options

@title Developer Setup
@group developer
How to configure a Phabricator development environment.
Overview
========
There are some options and workflows that may be useful if you are developing
or debugging Phabricator.
Configuration
=============
To adjust Phabricator for development:
- Enable `phabricator.developer-mode` to enable some options and show
more debugging information.
- Enable `phabricator.show-prototypes` to show all the incomplete
applications.
- See @{article: Using DarkConsole} for instructions on enabling the
debugging console.
Error Handling
==============
Errors normally go to DarkConsole (if enabled) and the webserver error log,
which is often located somewhere like `/var/log/apache/error_log`. This file
often contains relevant information after you encounter an error.
When debugging, you can print information to the error log with `phlog(...)`.
You can `phlog(new Exception(...))` to get a stack trace.
You can print information to the UI with `throw new Exception(...)`,
`print_r(...)`, or `var_dump(...)`.
You can abort execution with `die(...)` if you want to make sure execution
does not make it past some point. Normally `throw` does this too, but callers
can `catch` exceptions; they can not catch `die(...)`.
Utilities
=========
After adding, renaming, or moving classes, run `arc liberate` to rebuild
the class map:
```
phabricator/ $ arc liberate
```
Until you do this, Phabricator won't recognize your new, moved, or renamed
classes. You do not need to run this after modifying an existing class.
After any modifications to static resources (CSS / JS) but before sending
changes for review or pushing them to the remote, run `bin/celerity map`:
```
phabricator/ $ ./bin/celerity map
```
This rebuilds the static resource map.
If you forget to run these commands you'll normally be warned by unit tests,
but knowing about them may prevent confusion before you hit the warnings.
Command Line
============
Almost every script supports a `--trace` flag, which prints out service
calls and more detailed error information. This is often the best way to get
started with debugging command-line scripts.
Performance
===========
Although it is more user-focused than developer-focused, the
@{article:Troubleshooting Performance Problems} guide has useful information
on the tools available for diagnosing and understanding performance problems.
Custom Domains
==============
If you're working with applications that support custom domains (like Phurl or
Phame) you can normally test them by adding more entries to your webserver
configuration that look exactly like the primary entry (or expanding the
primary entry to match more domains).
Phabricator routes all requests based on host headers, so alternate domains
do not normally need any kind of special configuration.
You may also need to add `/etc/hosts` entries for the domains themselves.
Creating Test Data
==================
You can create test objects with the "Lipsum" utility:
```
phabricator/ $ ./bin/lipsum help generate
phabricator/ $ ./bin/lipsum generate ...
```
Test data can make your local install feel a little more realistic. With
`--quickly`, you can generate a large amount of test data to help test issues
with performance or scale.
c4science · Help