History

Torsten Ruger b9b05378f5 split the if test into a plain and one with call		2016-12-27 20:32:38 +02:00
..
arm	memory instruction refactor (small)	2016-12-14 21:05:24 +02:00
bench	remove references to soml	2016-12-11 12:55:03 +02:00
elf	much less test code (s-exps) needed because compiler defaults to main	2016-12-17 00:19:23 +02:00
lib	fix silly scope	2016-12-15 22:28:42 +02:00
melon	bringing the locals into the ruby method	2016-12-21 11:30:35 +02:00
register	split the if test into a plain and one with call	2016-12-27 20:32:38 +02:00
typed	rename also get_slot, to slot_to_reg	2016-12-25 18:05:39 +02:00
helper.rb	renames compiler to method_compiler	2016-12-18 14:15:19 +02:00
README.md	add missing tests (that were not liked to the travis script)	2016-12-08 12:55:16 +02:00
test_all.rb	folded salama-arm in	2016-12-14 13:43:13 +02:00

README.md

Testing

Testing is off course great, and well practised in the ruby community. Good tests exists in the parts where functionality is clear: Parsing and binary generation.

But it is difficult to write tests when you don't know what the functionality is. Also TDD does not really help as it assumes you know what you're doing.

I used minitest as the framework, just because it is lighter and thus when the time comes to move to salama, less work.

All

'''' ruby test/test_all.rb ''''

Parfait

Well, test Parfait. Not perfect, but growing as bugs appear. Basics are ok though.

Compiler

Different kinds of quite minimal tests that ensure we can go from parsed to code.

Fragments

Much more elaborate tests of the compling functionality. All code constructs and their output in terms of instructions are tested.

vm

Mostly tests about the Parfait compatibility layer and padding (for assmenbly). Slightly bad name ... wip