Differences

This shows you the differences between two versions of the page.

--- eiffel:etf:start [2020/02/05 02:11]
jonathan
+++ eiffel:etf:start [2020/03/08 21:25] (current)
jonathan
@@ Line 3: / Line 3: @@
 ======= Eiffel Testing Framework (ETF) =======
+ETF was developed by the Software Engineering Laboratory at York University -- to introduce students to open-ended design problems, while allowing instructors to check the feasibility and correctness of the designs via acceptance testing. See [[https://ieeexplore.ieee.org/document/8501488|2018 IEEE 8th International Model-Driven Requirements Engineering Workshop (MoDRE)]].
-Here is a [[https://www.youtube.com/playlist?list=PL5dxAmCmjv_5Xnqm0HcCBe6B4LB-75peL|series of tutorial videos]] to get you started on the Eiffel Testing Framework (ETF). Updated videos: [[https://www.youtube.com/playlist?list=PL5dxAmCmjv_5unIgLB9XiLwBey105y3kI| ETF Video Tutorials]].
-ETF adds the ability to introduce user requirements into a seamless process of software construction at a point before concrete user interfaces can be specified. A user writes an input grammar (input.txt) that describes all possible user input operations. The etf command operating on the input grammar generates code that takes care of parsing input commands. The software designer can then focus on developing the business logic (the model).
+  * [[https://www.youtube.com/playlist?list=PL5dxAmCmjv_5unIgLB9XiLwBey105y3kI| ETF Video Tutorials]].
-ETF allows software developers to produce use cases that can be turned into acceptance tests, and that then free the developer to develop the business logic while not losing sight of the user requirements. This allows requirements to become part of a seamless development from requirements to implemented code, and allowing change even at the level of requirements.
+  * For a small introductory step-by-step tutorial (can be done in less than 30 minutes), see [[https://github.com/yuselg/ETF|Designing a simple bank app with ETF]].
-===== Getting Started with ETF =====
+  * For Regression Testing see [[https://youtu.be/KdlZfspZjn8|Regression Testing Video]]. Also shows how to use Github if you re working on your Laptop or at home.
-(1) You need a file definitions.txt, for example:
+  * [[.debugger:|Using the debugger with an ETF acceptance test]]
-<code>
-system bank
-type NAME = STRING
-new(name: NAME)
-  -- create a new bank account
-deposit(name: NAME; amount: VALUE)
-  -- deposit "amount" into the account
-withdraw(name: NAME; amount: VALUE)
-  -- withdraw "amount" from the account
-transfer(name1: NAME; name2: NAME; amount: VALUE)
-  -- transfer `amount' from `name1' to `name2'
-</code>
-(2) You need an acceptance test file, e.g.. at1.txt
-<code>
-new("Steve")
-deposit("Steve", 10)
-deposit("Steve", 5)
-withdraw("Steve", 10)
-</code>
-(3) On a Prism workstation (or SEL-VM), execute the etf command to generate starter code for the bank system
-<code>
-mkdir src
-ls
-</code>
-You should see
-<code>
-at1.txt  definitions.txt  src/
-</code>
-Now do:
-<code>
-etf -new definitions.txt src/
-</code>
-The above command will generate the starting code in directory src/.
-Create a temporary directory in which to compile the EIFGENs (not needed on your Laptop)
-<code>
-mkdir /tmp/$USER
-</code>
-(4) To open the generated project in EiffelStudio:
-<code>
-estudio src/bank.ecf &
-</code>
-The above command will compile the project using EStudio. There is a long compile cycle at this point, so be patient. To speed up the compile cycle use the /tmp directory as follows
-{{:etf:eifgen.png?400|}}
-(Assuming the login is "student", i.e. $USER. Use your Prism login instead of "student".). This will ensure that the EIFGENs are located on the current workstation in its /tmp/student directory, and thus avoid having to swap the many C files and compilations over the slow network to the NFS server, and help to speed up compilation.
-After compilation, if you execute the system, you get the Graphical User Interface:
-{{:etf:gui.png?600|}}
-(5) We would like to run acceptance tests (such as at1.txt) in batch mode from the command line. To do this, change a switch in the ROOT class:
-{{:etf:command.png?600|}}
-(6) Create a symbolic link to the workbench executable
-<code>
-ln -s /tmp/student/EIFGENs/ehealth/W_code/bank bank.exe
-</code>
-(7) You can now execute the acceptance test
-<code>
-bank.exe -b at1.tx
-</code>
-to obtain
-<code>
-System State: default model state (0)
-->new("Steve")
-  System State: default model state (1)
-->deposit("Steve",10)
-  System State: default model state (2)
-->deposit("Steve",5)
-  System State: default model state (3)
-->withdraw("Steve",10)
-  System State: default model state (4)
-</code>
-====== Testing using the Workbench Code ======
-==W_code (Workbench)==
-<code>
-ec -c_compile -project_path /tmp/$USER -config system.ecf
-</code>
-You can create a symbolic link to the W_code executable:
-<code>
-ln -s /tmp/jonathan/EIFGENs/ehealth/W_code/system system.exe
-</code>
-You can now run a batch test as:
-<code>
-./system.exe -b at1.txt
-</code>
-If you recompile the code then you can execute again to see the results. You can also use the EStudio IDE to set break points and debug your code:
-If you use the Haskell testing script, you could automatically test and get results such as:
-<code>
-indigo 189 % ./etf_test.sh
-Building executable from ETF_Test.
-...
-=========================
-Test Results: 6/6 passed.
-=========================
-Success: ./log/at1.expected.txt and ./log/at1.actual.txt are identical.
-Success: ./log/at2.expected.txt and ./log/at2.actual.txt are identical.
-Success: ./log/at10.expected.txt and ./log/at10.actual.txt are identical.
-Success: ./log/at11.expected.txt and ./log/at11.actual.txt are identical.
-Success: ./log/at12.expected.txt and ./log/at12.actual.txt are identical.
-Success: ./log/at13.expected.txt and ./log/at13.actual.txt are identical.
-=========================
-Test Results: 6/6 passed.
-=========================
-</code>
-==F_code (Finalized)==
-<code>
-ec  -c_compile -finalize -project_path /tmp/$USER -config system.ecf
-</code>
-====== Useful links ======
-  * [[https://wiki.eecs.yorku.ca/project/sel-students/p:tutorials:etf:script:start|Haskell Script]] to check actual vs. expected acceptance tests.
-  * Using [[https://wiki.eecs.yorku.ca/project/sel-students/p:tutorials:etf:debugger:start|EStudio debugger]] to step through ETF code
-  * Run an ETF console-app as a [[https://wiki.eecs.yorku.ca/project/sel-students/p:tutorials:etf:web:start|web-app]]
-====== Adding Subclusters in ETF ======
-In ETF, the developer places the business logic in the sub-cluster ''model''. This sub-cluster is **recursive**. To add a sub-cluster (e.g. ''operations'') to ''model'', use the operating system to create a new folder under model with the name "operations".
-  model
-  ├── etf_model.e
-  ├── etf_model_access.e
-  └── operation
-      └── operation.e
-Then, in the IDE, recompile. The sub-cluster operations will automatically be there. There is no need to use the Settings tool of the IDE to add a new cluster (in fact, this will only create problems in this recursive case).
-===== Tests cluster =====
-A ''tests'' folder is also automatically generated. However, this folder is not automatically part of the system. To add this folder as a cluster, use the Settings tool of the IDE. Right click on the __Clusters__ group, and then do as follows:
-{{:etf:tests.png|}}
-Then recompile. Cluster "tests" will now show.

Software Engineering Documentation

User Tools

Site Tools

Differences

Page Tools