[clang] [PS4,PS5][Driver] Check for absent SDK when -nostdlib/-nodefaultlibs (PR #107112)

[Edd Dawson via cfe-commits]

================
@@ -36,9 +35,6 @@
 // RUN: env SCE_ORBIS_SDK_DIR=.. %clang -Winvalid-or-nonexistent-directory -### -emit-ast -isysroot foo -target x86_64-scei-ps4 %s 2>&1 | FileCheck -check-prefix=WARN-ISYSROOT -check-prefix=NO-WARN %s
 // RUN: env SCE_ORBIS_SDK_DIR=.. %clang -Winvalid-or-nonexistent-directory -### --sysroot=foo/ -isysroot foo -target x86_64-scei-ps4 %s 2>&1 | FileCheck -check-prefix=WARN-ISYSROOT -check-prefix=NO-WARN %s
 
-// RUN: env SCE_ORBIS_SDK_DIR=.. %clang -Winvalid-or-nonexistent-directory -### -nostdlib -target x86_64-scei-ps4 %s 2>&1 | FileCheck -check-prefix=WARN-SYS-HEADERS -check-prefix=NO-WARN %s
-// RUN: env SCE_ORBIS_SDK_DIR=.. %clang -Winvalid-or-nonexistent-directory -### -nodefaultlibs -target x86_64-scei-ps4 %s 2>&1 | FileCheck -check-prefix=WARN-SYS-HEADERS -check-prefix=NO-WARN %s
----------------
playstation-edd wrote:

I generally favaour leaving the test as it would be if the code was like that in the first place, as it avoids indefinite accumulation. Assuming the commit goes in as thing stands, there's no more reason to check for interaction with `-nodefaultlibs` as there is for interaction with `-nolibc` (which the code ignores, currently). And so it also avoids hitches for the reader ("well _of course_ there's no interaction with xyz... why would be checking it?").

With any luck, near future changes will go on to implement complete behaviour for all these `-no...` switches for library control. At which point new affirmative checks will be introduced.

With my case made, I'm also content to amend if we want that.

https://github.com/llvm/llvm-project/pull/107112