summaryrefslogtreecommitdiffstats
path: root/src/plugins/platforms/wasm/qwasmdrag.cpp
diff options
context:
space:
mode:
authorMarc Mutz <[email protected]>2022-12-06 12:56:15 +0100
committerMarc Mutz <[email protected]>2022-12-22 11:16:02 +0000
commit8420d3e0b205d0ef2367224afd3040c68e28bcc9 (patch)
treeb624188448264caff82fc564cd17ca0f839d81ed /src/plugins/platforms/wasm/qwasmdrag.cpp
parent115b1eab86d0395dc29f90d271bfdd29d425442e (diff)
tst_QString: test with QT_RESTRICTED_CAST_FROM_ASCII, too
QtBase didn't contain any checks for QT_RESTRICTED_CAST_FROM_ASCII, so a recent addition to the QString::append/insert/prepend overload set made calls with C string literal arguments ambiguous without the CI noticing. We had a similar problem with QString::multiArg. To increase test coverage, we now run tst_qstring two times: - without any define - with QT_RESTRICTED_CAST_FROM_ASCII (lots of changes necessary) Most removals are expected, because they disable tests that check the implicit conversions from QByteArray and const char*, but the relational operators with QLatin1String objects might warrant fixing. In some places, when the conversion wasn't the functionality under test, replaced C string literals or QByteArrays with QLatin1String. We should also test with QT_NO_CAST_FROM_ASCII, but that's even larger surgery. QString doesn't have a ctor from std::nullptr_t, so QString s = nullptr; doesn't compile in C++17 mode, but does in C++20 mode, due to the const char8_t* ctor. Pick-to: 6.5 6.4 6.2 5.15 Change-Id: I0c5a31719a4b8dd585dd748e0ca0d99964866064 Reviewed-by: Alexey Edelev <[email protected]> Reviewed-by: hjk <[email protected]> Reviewed-by: MÃ¥rten Nordheim <[email protected]>
Diffstat (limited to 'src/plugins/platforms/wasm/qwasmdrag.cpp')
0 files changed, 0 insertions, 0 deletions