izaya/collapseos
1
1
Fork 0
Code Issues Pull Requests Releases Wiki Activity
217df20d77
collapseos/tests
History
Virgil Dupras 48078d9c9c forth: Replace "SKIP?" with "(?br)" 
There is an alternate git history where I continued the Forth-ification of
words, including "SKIP?", but that was a bad idea: because that word was
written by flow control immediates, I stepped into quicksands where stability
became necessary in z80c.fs and I couldn't gracefully get out of it.

I'm stepping back and take this opportunity to replace the shoddy SKIP? algo
with a more straightforward (?br) implementation.

(br) and (?br) will always stay in boot code where it's easier manage a stable
ABI.
2020-03-29 09:10:23 -04:00
..
avra recipes/sms/kbd: use Collapse OS' AVR assembler 2020-02-26 23:10:43 -05:00
forth forth: Replace "SKIP?" with "(?br)" 2020-03-29 09:10:23 -04:00
shell tests/shell: bring back the static test.cfs 2019-12-31 15:17:22 -05:00
unit zasm: allow zasm to omit its 3rd argument 2020-02-18 15:46:55 -05:00
zasm emul/zasm: use libcfs 2019-12-31 15:07:39 -05:00
Makefile tests: add Forth tests 2020-03-19 15:43:48 -04:00
README.md

README.md

Testing Collapse OS

This folder contains Collapse OS' automated testing suite. To run, it needs tools/emul to be built. You can run all tests with make.

zasm

This folder tests zasm's assembling capabilities by assembling test source files and compare the results with expected binaries. These binaries used to be tested with a golden standard assembler, scas, but at some point compatibility with scas was broken, so we test against previously generated binaries, making those tests essentially regression tests.

Those reference binaries sometimes change, especially when we update code in core libraries because some tests include them. In this case, we have to update binaries to the new expected value by being extra careful not to introduce a regression in test references.

unit

Those tests target specific routines to test and test them using tools/emul/runbin which:

  1. Loads the specified binary
  2. Runs it until it halts
  3. Verifies that A is zero. If it's not, we're in error and we display the value of A.

Test source code has no harnessing and is written in a very "hands on" approach. At the moment, debugging a test failure is a bit tricky because the error code often doesn't tell us much.

The convention is to keep a testNum counter variable around and call nexttest after each success so that we can easily have an idea of where we fail.

Then, if you need to debug the cause of a failure, well, you're on your own. However, there are tricks.

  1. Run unit/runtests.sh <name of file to test> to target a specific test unit.
  2. Insert a halt to see the value of A at any given moment: it will be your reported error code (if 0, runbin will report a success).

shell

Those tests are in the form of shell "replay" files. Every ".replay" file in this folder contains the contents to type in the shell. That contents is piped through the shell and the output is then compared with the corresponding ".expected" file. If they match exactly, the test passes.