Friday, April 22, 2005 6:17 PM (some stuff can be obsolete) ============================== Below are enumerated all the ASL operators (32) which for now have no separate tests. 1. Intensively applied in other tests and thus might be considered being already verified. Do simple separate test for each of these operators which will be the places where, in the future, the additional sub-tests for these operators, if any, could be located. ArgX LocalX Name Buffer String Store CopyObject 2. Indirectly verified Include 3. There is no usual way to verify them. The tests might be either omitted or could be done basing on the extension of the "memory consumption instrumentation" supplying the ASL tests with the required internal information. Fatal Notify 4. Can't be verified in our test suite Load Unload External LoadTable BreakPoint DefinitionBlock (each test is represented as this kind block) DataTableRegion 5. It is not worthy of to be definitely verified. Do simplest test, check only that execution of this operator doesn't produce exception. 17.5.83 NoOp Description: This operation has no effect. To verify the sentence like the above definitely is too complicated thing - if you have proved it for N conditions it nevertheless can fail in (N+1)-th condition. The relevant test should involve all the possible functionality of the interpreter inserting everywhere this NoOp and verifying everything each time. Apparently, it is not worth the trouble. It is a simple operator and it can be and should be implemented with the care once. NoOp 6. Can in principle be ever in the future verified (applying auxiliary script) (see e-mail RE:Planning ACPICA validation 19.11.2004, 303_test_for_debug attachment for more details) Debug 7. Do the simple test, but not comprehensive The comprehensive verification of this operator could be conceived based on the following principle - do replacement of all (or most of them, or at least globally declared of them) the objects applied in the test suite by the Aliases to them. This could be done basing on the conditional #if ASL compilation which is not provided now by iASL, and require the separate compilation and run of the test suite. Such the comprehensive test would actually show that "AliasObject acts exactly the same as SourceObject" (as far as the test suite actually provides coverage). But we have at least one reason for the simple test, but not comprehensive - there is no #if ASL ability, and the second issue is either the verification of this operator is worthy of the considerable and full of work restructurization of the test suite. We propose to do the simple test which use only ObjectType (and SizeOf) for verification of each type of SourceObject. One more checking will be the overwritings with the different type object: CopyObject(XXX, AliasObject) --> check SourceObject CopyObject(XXX, SourceObject) --> check AliasObject This operator doesn't add functionality, it is only for convenience (but since provided it should be verified). Alias 8. Require the new tests designing and implementing Return Method Function Field BankField IndexField OperationRegion Scope Device Processor ThermalZone PowerResource