tests: accept %p printing in uppercase and without leading 0x

[mguetschow]

Contribution description

The output of the specifier %p is implementation defined, and mpaland-printf gives something different than newlib. Be less strict in what the tests expect.

Testing procedure

Build and run any of the changed tests on a board now pulling mpaland-printf. Notice that they fail on master and not with these changes.

Issues/PRs references

Should fix the failing nightlies after #21438

[maribu]

This will still not pass with mpaland-printf, as it also uses "(nil)" like glibc does on NULL.

The following would pass:

diff --git a/tests/core/msg_queue_print/tests/01-run.py b/tests/core/msg_queue_print/tests/01-run.py
index a03118d268..f78b50870e 100755
--- a/tests/core/msg_queue_print/tests/01-run.py
+++ b/tests/core/msg_queue_print/tests/01-run.py
@@ -41,12 +41,12 @@ def expect_some(child, size, avail, range_start):
 
 def expect_content(child, counter):
     if counter == 0:
-        if os.environ.get('BOARD') in ['native', 'native32', 'native64']:
-            child.expect_exact('type: 0x0000, content: 0 ((nil))')
-        else:
-            child.expect(r'type: 0x0000, content: 0 \((0x)?0+\)')
+        child.expect(r'type: 0x0000, content: 0 \((\(nil\)|0x0+\))')
     else:
-        child.expect_exact(f'type: 0x{counter:04x}, content: {counter} (0x{counter:x})')
+        child.expect(r'type: ((0x)?[a-fA-F0-9]+), content: (\d+) \(((0x)?[a-fA-Z0-9]+)\)')
+        assert int(child.match.group(1), 16) == counter, f"Expected {counter:x} as type, got child.match.group(1)"
+        assert int(child.match.group(2)) == counter, f"Expected {counter} as content, got child.match.group(2)"
+        assert int(child.match.group(3), 16) == counter, f"Expected {counter:x} as ptr, got child.match.group(3)"
 
 
 def testfunc(child):

[mguetschow]

I think that's only true after applying the patch in #21678

[maribu]

Indeed. I updated the patch above, I think would now work for all ways to do %p.

[riot-ci]

Murdock results

✔️ PASSED

4767b98 tests/core/msg_queue_print: accept (nil) instead of 0x0+ for all boards

Success	Failures	Total	Runtime
189	0	189	03m:11s

Artifacts

• Documentation preview

[mguetschow]

Ready from my side, should still fix the remaining failing nightly.