44 lines
1.8 KiB
Plaintext
44 lines
1.8 KiB
Plaintext
|
TODO
|
||
|
|
====
|
||
|
|
Todo tasks can be found in the bug tracker at http://easejs.org/bugs. Below is
|
||
|
|
a list of the glaring issues that may be found at first glance.
|
||
|
|
|
||
|
|
|
||
|
|
COUPLING
|
||
|
|
--------
|
||
|
|
During the beginning of the project, all modules were represented as object
|
||
|
|
literals, which is similar to the concept of a Singleton. This satisfied the
|
||
|
|
needs of the early project, but it soon evolved far past its original design and
|
||
|
|
resulted in a tightly coupled system that was difficult to maintain, add to and
|
||
|
|
test. Refactoring into prototypes is ongoing.
|
||
|
|
|
||
|
|
|
||
|
|
UNIT TESTS
|
||
|
|
----------
|
||
|
|
The unit tests are currently split between a few different formats, depending on
|
||
|
|
filename and test construction.
|
||
|
|
|
||
|
|
- Tests created during the beginning of the project have a test-* prefix. These
|
||
|
|
are more likely to represent integration or system tests rather than unit
|
||
|
|
tests. They should be converted to files with a *Test.js suffix and be
|
||
|
|
refactored to test only the unit.
|
||
|
|
- Newer tests use require( 'common' ).testCase() in order to run the tests. The
|
||
|
|
older format separates tests with self-executing functions. The former should
|
||
|
|
be used.
|
||
|
|
|
||
|
|
When running the tests, you will notice that the older tests output nothing
|
||
|
|
whereas the newer tests output the status of each test in addition to a summary
|
||
|
|
at the end of the test case. Ultimately, a test runner will defer final
|
||
|
|
statistics until the end of all running tests. The Makefile will need to be
|
||
|
|
modified. This sacrifices the ability to use -jN with the tests, but will
|
||
|
|
ultimately speed up the tests by eliminating the need to run a separate process
|
||
|
|
for each test case (the performance benefit can be seen by running the tests in
|
||
|
|
the browser).
|
||
|
|
|
||
|
|
|
||
|
|
PERFORMANCE TESTS
|
||
|
|
-----------------
|
||
|
|
Performance tests need to be written for every aspect of the system. They will
|
||
|
|
ultimately be graphed to show the relative performance across versions of the
|
||
|
|
software.
|