Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/rushilp2311/eecs3311_lab3
https://github.com/rushilp2311/eecs3311_lab3
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/rushilp2311/eecs3311_lab3
- Owner: rushilp2311
- Created: 2020-10-31T23:22:42.000Z (about 4 years ago)
- Default Branch: master
- Last Pushed: 2020-11-06T03:52:20.000Z (about 4 years ago)
- Last Synced: 2024-11-05T14:40:38.477Z (3 months ago)
- Language: Eiffel
- Size: 4.86 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.txt
Awesome Lists containing this project
README
0. See this tutorial video to get started:
https://www.youtube.com/watch?v=Do_dpN66Jtk&list=PL5dxAmCmjv_5q9wBFXV-M4S4VbB1T0EJg&index=8
1. Upload the starter `space_defender_1.zip` to your Prism account (e.g., via an FTP program).
Unzip it: unzip space_defender_1.zip2. Login to your Prism account, change the current directory to this starter folder, then run the following command:
etf -new space_defender_1.grammar.txt .
3. Do the following re-arrangements of files:
A. Move the given acceptance tests to the `tests/acceptance/instructor` folder:at001.txt
at001.expected.txt
at002.txt
at002.expected.txt
at003.txt
at003.expected.txt
B. Move the following files to the `docs` folder:
messages.txt
OOSC2_UNDO.pdf
space_defender_1.grammar.txt
4. Change the root class:A. Run: gedit root/root.e &
B. Change the `switch` query so that `eft_cl_show_history` is uncommented:
feature -- Queries
switch: INTEGER
-- Running mode of ETF application
do
-- Result := etf_gui_show_history -- GUI mode
Result := etf_cl_show_history
-- Result := unit_test -- Unit Testing mode
end5. Compile the project:
ec19.05 -c_compile -freeze -config space_defender_1.ecf
6. Change the current directory to: `regression-testing`.
7. Invoke the regression testing script: ./ETF_Test.py
8. Then you should see a message saying that NONE of the three starter tests passed.
Actual outputs produced by your program are placed: log/instructor/at001.actual.txt
Expected outputs produced by the oracle are placed: log/instructor/at001.expected.txt
Use `meld` to see the difference between the actual and expected outputs. e.g.,
meld log/instructor/at001.actual.txt log/instructor/at001.expected.txt &9. How to go from here?
A. Complete your development in "space_defender_1/abstract_ui/user_commands" and "space_defender_1/model".
B. Add more test cases to "tests/acceptance/student".
C. Constantly re-run the regression testing script (this is how you will be graded).
D. Submit the intermediate version of your work constantly.